Summit Program

All sessions will be held in Riverview B unless otherwise noted.

 

Thursday, June 19, 2014

8:00 a.m.–9:00 a.m. Thursday

Continental Breakfast

Columbus Foyer

9:00 a.m.–9:15 a.m. Thursday

Opening Remarks

Program Chair: Chris St. Pierre, Metacloud

9:15 a.m.–10:30 a.m. Thursday

Keynote Address

Things Will Change

Eric Windisch, Docker

Eric will cover a spectrum from today's tools to what he see as the coming evolution of those tools as they converge, and as we approach an Internet of Things. Practically, this will draw upon his experiences and the implementations of Chef, Docker, and Heat.

Eric will cover a spectrum from today's tools to what he see as the coming evolution of those tools as they converge, and as we approach an Internet of Things. Practically, this will draw upon his experiences and the implementations of Chef, Docker, and Heat.

Available Media
10:30 a.m.–11:00 a.m. Thursday

Break with Refreshments

Columbus Foyer

11:00 a.m.–11:45 a.m. Thursday

Invited Talk

Continuous Integration and Collaboration in Managed Environments

Joe Netzel, CFEngine

Joe Netzel spent more than 15 years in the trenches as a System Administrator for IBM, General Electric, and Lockheed Martin before joining CFEngine in 2011.

One of the great challenges faced by large IT organizations revolves around the automation of desired state while allowing a diverse set of stakeholders the ability to provide direct input into the configuration management process. The traditional approach of well-defined boundaries and role-based silos quickly becomes a liability in fluid and complex environments where stakeholder analysis, input, and ability to act are critical to business success. Recently, the principles of Continuous Integration have been successfully applied to standard operational models by combining configuration management tools with version control and software review tools. By combining these tools with light-weight processes and a common CI methodology across stakeholder teams, the business needs of any given organization can be met in a more effective fashion.

One of the great challenges faced by large IT organizations revolves around the automation of desired state while allowing a diverse set of stakeholders the ability to provide direct input into the configuration management process. The traditional approach of well-defined boundaries and role-based silos quickly becomes a liability in fluid and complex environments where stakeholder analysis, input, and ability to act are critical to business success. Recently, the principles of Continuous Integration have been successfully applied to standard operational models by combining configuration management tools with version control and software review tools. By combining these tools with light-weight processes and a common CI methodology across stakeholder teams, the business needs of any given organization can be met in a more effective fashion.

In this talk, you will learn how to use Continuous Integration principles that will allow full stakeholder access to the creation and promotion of configuration management policies, from test to production, using CFEngine, Git, and Gerrit. In addition, lightweight process examples will be examined and incorporated into the collaboration flow.

Joe Netzel spent more than 15 years in the trenches as a System Administrator for IBM, General Electric, and Lockheed Martin before joining CFEngine in 2011.

Available Media

11:45 a.m.–12:30 p.m. Thursday

Invited Talk

Configuration as a Service

Nico Schottelius, ungleich GmbH

Nico Schottelius is a FOSS hacker and owner and CEO of the ungleich GmbH, a Swiss Linux/Unix infrastructure company. He grew up with Linux and FOSS and is developing the c-star (c*) software series, a set of highly productive and slim software packages.

He received his degree in information management in 2005 and a BsC in computer science in 2012.

Configuration Management is the de-facto standard to maintain infrastructures of all sizes. Even though some of the configuration management systems can bootstrap themselves, deploying and maintaining a configuration management system requires a permanent investment.

Computational resources have successfully been outsourced in the form of cloud computing and various other services like infrastructure (IaaS) or software (SaaS) are also available as a service. This talk discusses the possibility of moving from running a dedicated configuration management system to using configuration as a service, based on the cdist cloud configuration service, which allows users to configure their hosts without running a configuration management system on their own.

Configuration Management is the de-facto standard to maintain infrastructures of all sizes. Even though some of the configuration management systems can bootstrap themselves, deploying and maintaining a configuration management system requires a permanent investment.

Computational resources have successfully been outsourced in the form of cloud computing and various other services like infrastructure (IaaS) or software (SaaS) are also available as a service. This talk discusses the possibility of moving from running a dedicated configuration management system to using configuration as a service, based on the cdist cloud configuration service, which allows users to configure their hosts without running a configuration management system on their own.

Nico Schottelius is a FOSS hacker and owner and CEO of the ungleich GmbH, a Swiss Linux/Unix infrastructure company. He grew up with Linux and FOSS and is developing the c-star (c*) software series, a set of highly productive and slim software packages.

He received his degree in information management in 2005 and a BsC in computer science in 2012.

Available Media

12:30 p.m.–1:30 p.m. Thursday

FCW '14 Luncheon

Grand Ballroom ABC

1:30 p.m.–2:15 p.m. Thursday

Invited Talk

Automating Orchestration in the Cloud with Ubuntu Juju

Charles Butler, Canonical

Charles is a developer/admin from Pittsburgh, PA. He joined the Juju Solutions team at Canonical in January, 2014, and has an interest in service oriented architecture, and automation. His background stems from three years working in marketing managing small business clusters in the cloud, and leading a five-man development team.

Most of you have automated your deployments. But how does it stand up to the bus factor? How reusable is your automation solution? Do you find yourself wrestling more with the annoyances of scaling than deploying your actual services?

Service orchestration is now the easiest, most flexible, and most approachable solution to shipping your infrastructure woes to /dev/null. Starting with Ubuntu’s 14.04 LTS release, our suite of devops tools are baked right into the base Ubuntu Server installation.

In this session, we'll quickly go over the basic concepts of Juju and spend the rest of the time walking through explicit examples of Juju in action. We'll look at stacks of services managed by Juju and the charms behind those services.

We'll talk about how to get you up and running quickly with multi-sharded MongoDB clusters, Hadoop, Cassandra, and more in just a few commands. All ready to be deployed and reused to meet your needs.

Most of you have automated your deployments. But how does it stand up to the bus factor? How reusable is your automation solution? Do you find yourself wrestling more with the annoyances of scaling than deploying your actual services?

Service orchestration is now the easiest, most flexible, and most approachable solution to shipping your infrastructure woes to /dev/null. Starting with Ubuntu’s 14.04 LTS release, our suite of devops tools are baked right into the base Ubuntu Server installation.

In this session, we'll quickly go over the basic concepts of Juju and spend the rest of the time walking through explicit examples of Juju in action. We'll look at stacks of services managed by Juju and the charms behind those services.

We'll talk about how to get you up and running quickly with multi-sharded MongoDB clusters, Hadoop, Cassandra, and more in just a few commands. All ready to be deployed and reused to meet your needs.

Be ready with questions... What's the point? Which problems was it created to solve? What's it really look like? What about my existing Puppet/Chef infrastructure? What is a Juju charm anyway? Do I need to go to charm school to get started?

Charles is a developer/admin from Pittsburgh, PA. He joined the Juju Solutions team at Canonical in January, 2014, and has an interest in service oriented architecture, and automation. His background stems from three years working in marketing managing small business clusters in the cloud, and leading a five-man development team.

Available Media
2:15 p.m.–3:30 p.m. Thursday

Invited Talk

Software-Defined Infrastructure

Nick Cammorato, TERC

Nick Cammorato is a former software engineer who currently heads up the ops team for TERC, a non-profit educational research institution. He has over a decade of experience in system and network administration, most of which has been spent trying to automate himself out of a job.

Software-Defined Infrastructure, or SDI, is the management of your entire infrastructure as software. Modern provisioning techniques combined with configuration management have progressively blurred the line between software and infrastructure. While the complication of actually implementing it can vary immensely, it has been increasingly possible to abstract more and more into a single DSL.

Using a traditional infrastructure and simplified application stack as an example, we will examine the necessary components to reach an approximation of SDI, identify the brick walls you will run into, ways to work around them, the importance of separation of concerns and sources of truth, and discuss dependency resolution strategies and classification. Continuous Integration and Continuous Delivery (CI/CD) will be discussed briefly.

Software-Defined Infrastructure, or SDI, is the management of your entire infrastructure as software. Modern provisioning techniques combined with configuration management have progressively blurred the line between software and infrastructure. While the complication of actually implementing it can vary immensely, it has been increasingly possible to abstract more and more into a single DSL.

Using a traditional infrastructure and simplified application stack as an example, we will examine the necessary components to reach an approximation of SDI, identify the brick walls you will run into, ways to work around them, the importance of separation of concerns and sources of truth, and discuss dependency resolution strategies and classification. Continuous Integration and Continuous Delivery (CI/CD) will be discussed briefly.

Nick Cammorato is a former software engineer who currently heads up the ops team for TERC, a non-profit educational research institution. He has over a decade of experience in system and network administration, most of which has been spent trying to automate himself out of a job.

Available Media
3:30 p.m.–4:00 p.m. Thursday

Break with Refreshments

Columbus Foyer

4:00 p.m.–4:45 p.m. Thursday

Invited Talk

The Future of In-container Configuration Management

Diego Zamboni, CFEngine

Diego Zamboni is a computer scientist, consultant, author, programmer, and sysadmin who works as Senior Security Advisor and Product Manager at CFEngine. He has more than 20 years of experience in system administration and security and has worked in both the applied and theoretical sides of the computer science field. He holds a Ph.D. from Purdue University, has worked as a sysadmin at a supercomputer center, as a researcher at the IBM Zurich Research Lab, and as a consultant at HP Enterprise Services. He is the author of the book "Learning CFEngine 3," published by O'Reilly Media. He lives in Queretaro, Mexico, with his wife and two daughters.

IT has undergone many technology disruptions over the last 30 years, from bare metal to VMs, to cloud and most recently containers. Each generation is bringing along new value propositions and enabling new applications, while at the same time making some technologies obsolete or less relevant.

The promised value propositions of containers are around speed and portability. By taking a historical perspective in considering the new value propositions and current application of containers, we will look at what role we might expect configuration management to play in managing containers from the inside. We will also look at which qualities a configuration management tool needs to satisfy in order to fill the needs of the container generation.

IT has undergone many technology disruptions over the last 30 years, from bare metal to VMs, to cloud and most recently containers. Each generation is bringing along new value propositions and enabling new applications, while at the same time making some technologies obsolete or less relevant.

The promised value propositions of containers are around speed and portability. By taking a historical perspective in considering the new value propositions and current application of containers, we will look at what role we might expect configuration management to play in managing containers from the inside. We will also look at which qualities a configuration management tool needs to satisfy in order to fill the needs of the container generation.

Diego Zamboni is a computer scientist, consultant, author, programmer, and sysadmin who works as Senior Security Advisor and Product Manager at CFEngine. He has more than 20 years of experience in system administration and security and has worked in both the applied and theoretical sides of the computer science field. He holds a Ph.D. from Purdue University, has worked as a sysadmin at a supercomputer center, as a researcher at the IBM Zurich Research Lab, and as a consultant at HP Enterprise Services. He is the author of the book "Learning CFEngine 3," published by O'Reilly Media. He lives in Queretaro, Mexico, with his wife and two daughters.

Available Media

4:45 p.m.–5:30 p.m. Thursday

Invited Talk

Software Configuration Management at National Cancer Institute

Sarah Elkins, SRA International

Sarah Elkins has worked in configuration management for over 20 years, including the last 5+ at National Cancer Institute. Her CM experience includes teamwork and sole responsibility at large and small companies in multiple industries, including commercial software development as well as government contracting. She has developed and given internal CM training, and given internal and external presentations on various aspects of CM.

Configuration management at National Cancer Institute involves support for over 600 Web sites ranging from flat HTML to complex bioinformatics applications on a variety of container technologies. This 45-minute presentation will cover the SCM processes and technologies which enable software to move from source code repositories all the way to production servers at NCI. SCM activities have expanded within the last few years to include Git mirror as well as local Subversion repository hosting, Maven as well as Ant build support, build dependency artifacts management, and build and deploy automation setups for Windows as well as Linux. Details will address the promotion of sites/applications up the tiers to production, the services and tools we administer, and tool integration capabilities.

Configuration management at National Cancer Institute involves support for over 600 Web sites ranging from flat HTML to complex bioinformatics applications on a variety of container technologies. This 45-minute presentation will cover the SCM processes and technologies which enable software to move from source code repositories all the way to production servers at NCI. SCM activities have expanded within the last few years to include Git mirror as well as local Subversion repository hosting, Maven as well as Ant build support, build dependency artifacts management, and build and deploy automation setups for Windows as well as Linux. Details will address the promotion of sites/applications up the tiers to production, the services and tools we administer, and tool integration capabilities.

Sarah Elkins has worked in configuration management for over 20 years, including the last 6+ at National Cancer Institute. Her CM experience includes teamwork and sole responsibility at large and small companies in multiple industries, including commercial software development as well as government contracting. She has developed and given internal CM training, and given internal and external presentations on various aspects of CM.

Available Media