As you may recall, we’ve been recording a short series on the topic of “Agile Infrastructure” over on The Agile Executive podcast. Here is part two:

To listen to this podcast, download the podcast directly, subscribe to the blog/podcast feed in iTunes (or whatever), or click play below to hear it:

As part two in our (planned to be) three part or so series, Israel Gat, Andrew Shafer, and I get back together to discuss the idea of Agile Infrastructure. See part one for an overview of what “Agile Infrustructure” aims to be in the first place.This time, we talk about the difficulty of cultural change to make a more Agile IT process possible. We spend much time on “motherhood and apple-pie” topics as always happens when it comes to discussions of organizational change management, but drawing on our experiences in both small and very large companies, we start to pull apart the tactics for not only implementing a change to Agile IT, but coping with the friction that occurs during any change, esp. something as dramatic as changing to a more Agile way of delivering software and business services.

Here’s a feel for what’s in the episode:

• I start out by asking how organizations start the process of changing to more Agile IT operations.
• Andrew says that change starting at a grass roots level, bottom-up, and is far from wide-spread.
• Israel speaks to one anecdote with 20% from the top and 80% from the bottom.
• What’s the process for change? What’s the context for trying to get people to change?
• Andrew says that people have to care – they must be interested in doing more than passing the time and getting paid for it, as I put it. Ideally, champions can start to “manage upwards” too, as needed. Also, having an expert on Agile available for internal assistance and answering questions is key.
• Israel adds in his “Agile Executive” view. He speaks to using people’s desire to do better, but figuring out a window (or time and environment) in which it’ll succeed.
• Using small successes to build the room to do large successes.
• This discussion leads Andrew to remember a talk by Lisa Crispin where the testers and others began to understand the “business,” or the larger context they were operating in.
• I ask Andrew and Israel why it’s important for IT employees to rise above being someone who “just works here.”
• We then discuss how changing to the more role-fluid nature of Agile conflicts with the static nature of jobs in organizations, where people are assigned a specific role and aren’t expected to go outside that role.
• When it comes to knowing how well you’re doing, Andrew introduces the Dunning-Kruger effect, wherein people are bad at evaluating themselves, esp. when their context is limited to their own group or organization.
• Reflect and adapt… making sure you do this in Agile. Israel connects this back to building confidence and moral in the organization as a way to enable change and Agile. This relates to one of those group psychology studies that’s always fascinating, namely, people preferring confidence over expertise.

Disclosure: Reductive Labs is a client.

Download the episode directly here, subscribe to the RedMonk Radio podcast feed to have it automatically downloaded to iTunes or other podcatcher, or just click play below to listen to it right here:

In this sponsored episode of RedMonk Radio, I talk with Karl Pietri of Slideshare, along with Luke Kanies and Andrew Shafer of Reductive Labs. We discuss how Slideshare uses Puppet to manage all of it’s data-centers, spread across geographies, their own on-premise servers, and the cloud via Amazon EC2.

First, we go over the Slideshare architecture (a Ruby on Rails shop) and how that drives the layout of the data-center. Karl then tells us how Puppet is used to manage the different servers, highlighting how they use it for private/public cloud mixing. On this point, I ask Andrew to compare Slideshare’s use with others in the Puppet community.

Getting down the to the day-to-day operations, we spend the last part of the show talking about how Puppet fits into Slideshare’s release management (via SVN) and how it’s effected the relationship between development and operations.

Finally, we wrap up with a semantic-check on the word “pager”: are sys admins still running around with beepers still?

Disclosure: Reductive Labs is a client and sponsored this podcast.