Stop the Engineering Tug of War with an Open Source Office

What is an Open Source Office

Every commerical company has a engineering organization. Commerical engineering standard practice is to develop everything in private. However, times have changed and every company has some open source use and/or development going on. The use of open source without formal engineering support causes an unnecessary tug of war for resources. The solution is a group called the Open Source Office, dedicated to supporting the company’s open source work. An Open Source Office within a commercial organization, establishes the open source engineering support. A typical engineering organization focuses on vendors, operations, and possibility software development. Instead, the Open Source Office activities are focused on mitigating risks and supporting developers.

The detailed breakdown of the Open Source Office is into five components:

  1. Legal Oversight of Software
  2. Security Oversight of Software
  3. Commercial Support of Open Source Governance Organizations
  4. Sponsoring Open Source Projects and Tools
  5. Direct Support of Engineering Contributions to Open Source Projects

1. Legal Oversight of Software

Oversight is about risk mitigation. Every company has risk when using licensed software. Contracts determine how a company can use vendor privately licensed software. Open source licensed software requires a company’s legal department to set the standards of use. These standards of use are set by software license types.

Create a process for developers and managers to request a license be approved for use. Hire at least a part-time Open Source Legal Manager reporting directly to the Open Source Office, to respond to requests and own the process. Use a simple form with some basic questions of how, why, where, and when. Jira, SurveyMonkey, or Google Forms all work very well. Once a license is approved, publish the details where any developer can easily find it. is a good place.

So called permissive licenses like Apache 2.0 and MIT allow the software to be used with a minimum of restrictions and thus are favored. Other projects are using so called copy-left licenses that require more complicated rules. The benefit of the software will need to be balanced with the rules of use. You will need to work with your legal minds to figure out the software benefit risk balance for your company.

I recommend using the OSI and TODO organizations as resources. The Open Source Initiative (OSI) has an excellent reference plus mailing list discussions. The Linux Foundation TODO group has a legal workgroup that regularly discusses open source licensing. However, to participate in the TODO group, Linux Foundation membership is required.

Along with the approved open source software licenses, your developers need a clear process for joining, contributing, and creating open source projects. It is common practice to call using an existing open source based software project as Incoming (to your company) and contributing or creating a software project as Outgoing (to your company.) I recommend that once a license has been approved, that any contributions to open source project using the license be blanket approved by legal. You want to remove complexity and overhead.

Clearly document your Inbound and Outbound open source use policies at I recommend my simple open source IO policies below:

Inbound Open Source Software policy

  1. Working with a manager, find an open source project you want to use for company related work
  2. Verify that the software license is approved by legal here
  3. If the license if not approved or disapproved, request approval through the new license request at Legal will respond within a few days and start working with you on your request.
  4. If the software license is approved, review the use restrictions, if any. For questions, request answers through Legal will respond within a few days and start working with you on your request.
  5. Join or create your open source project to the software index at Developers, users, and managers will use this information for training, mentoring, and hiring.

Outbound Open Source Software policy

  1. Based on your own interests or working with a manager, find an open source project you want to contribute to. You can use as a guide for existing users and contributors within the company. These people can help you with training and mentoring.
  2. Once you find the open source project you want to contribute to, follow the contribution guidelines at OR
  3. If the project doesn’t yet exist at, then you need to verify or request the open source software license is approved here Once the open source software license is approved, then add the project to

2. Security Oversight of Software

Your legal department sets the software license standards of use. Your security department is tasked with enforcing compliance of those standards. Where it comes to using software, security also wants to reduce the risk of introducing software vulnerabilities. Security covers this for privately licensed software through contract language and penalties. For open source licensed software, your security department needs a software scanner to verify license compliance, viruses, and other security risks are identified before use.

Many organizations use Checkmarx, AppScan, or Blackduck by manually scanning a code repository for security and license compliance. While this accomplishes a one time need to verify the code is compliant, it ignores any ongoing risk. It is a better practice to include the scanning activity as part of your software pipeline (also called Continuous Integration). By adding the software scanning to the software pipeline, any problems can be triaged by the developers during the development cycle, rather than waiting until the last minute. You will need at least a part-time Open Source Security Manager reporting to the Open Source Office, to manage the process and to respond to requests.

Well documented and understood legal, security policies and processes is good governance. Knowing the rules through good governance makes for happy employees

3. Commercial Support of Open Source Governance Organizations

Open source does have overhead. Since open source projects are a public activity, not directly managed by business needs, you need some good governance to help organize and collaborate. Companies using open source need to support the organizations that support the open source community as a whole. These include OSI, OIN, Linux Foundation, OSCON, and OCP to name a few. Supporting these organizations with sponsorships allow volunteers like myself support open source public policy initiatives and mentor new open source leaders on best practices. Hire an VP Open Source to manage the office, personnel, and the governance relationships. It is important that the Open Source Office leader is at a senior level so the policies the resulting policies are followed.

On a side note, commercial open source is a term that has been applied to organizations like OpenStack that are a mostly corporate funded development. This term still works here as I am using it, as I am applying to the need of commercial open source support beyond just the developement projects, but also to the governance organizations that support those development projects. We as open source leaders, need to do a better job of highlighting the requirement of proper governance for success.

4. Sponsoring Open Source Projects and Tools

Beyond developer code contributions, your company needs to sponsor your open source project events as well. As your company supports open source governance through organizations like the Linux Foundation, your public development projects need a similar type of support. At least each open source project will need to get its contributors face-to-face around release time. Planning for the next release is critical to get be completed very close to the release of the current development work.

These events need to be planned, organized, and executed. Events with more than a handful of people over a few days, becomes a complicated, expensive enterprise. To fully support an open source project, you’re contributing company needs to source developers and event sponsorships. Hire at least a part-time Events Manager to be responsible for these activities. A person with marketing experience for the role of Events Manager is a likely fit. Make sure that the Events Manager works closely with the development community, so the events meet the developer’s needs.

Public facing tools require support and management just like the projects. GitHub and other components of the public project Continuous Integration need continuous management and contributions to keep the project development momentum going strong. What is often missed, is the administrative function of keeping the public repositories managed. All too often, an organization’s public GitHub repositories become littered with stale projects that are no longer under active development or supported. This hurts the company reputation, making other public project participation more difficult.

The Open Source Office should not only take administrative ownership of the GitHub organizations your company owns, but take an active role in curating the projects there. If a project is stale, then update the root readme to say the project is not under active development. A GitHub organization that holds stale projects, if there are many, can help as well. The project can always be moved again if it becomes active. Hire a Community Manager to administer the public GitHub organization(s).

5. Direct Support of Engineering Contributions to Open Source Projects

In addition to GitHub, there are many Continuous Integration tools such as Jenkins and Gerrit that can used by public projects. For your public projects, whatever the Continuous Integration tools are, support them as strongly as GitHub. By relying on others to support the infrastructure that make the public project successful, you are outsourcing a critical function. Get involved at least part-time in the project CI infrastructure to ensure no surprises. Your Community Manager will monitor the state of your public project momentum through the project CI infrastructure. As problems arise, the Community Manager will jump in and work to fix the problem or gather people to fix the problem.

Project Data Analytics

As your team contributes, your engineering leadership needs information for scrum retrospectives and employee reviews. Code commits only tell a small part of the project story. Project data analytics fills in the gaps of project activities like patch review quality, time spent on contributions, and patch set history. Understanding these types of developer analytics are critical for any project. Developer analytics on public projects are essential. Development tools like Jira and Aha have some basic reports, but are missing the detailed analysis that can help mentor a junior developer or highlight feature development momentum. There are tools that can fill this gap. I know of three so far, Stackalytics, Biterga and GitPrime Biterga and GitPrime can work on public and private repositories. I have been using Stackalytics for years, which is specific to OpenStack projects. It doesn’t have everything I would want. Stackalytics does allow for browsing patch review quality and filtered release contributions. Developer analytics has been a very important tool supporting open source projects.

Open Source Developers as Part of your Company Sprints

We have explained all the overhead for open source projects, except the most overlooked and arguably, the most important. Expand your developer support by supporting the public project development process within your internal development teams. To make this happen, your Open Source Office leadership needs to have authority over the product and project management. Dotted line management can work, while making sure open source is part of the annual reviews.

Day-to-day, week to week involvement with your internal engineering teams is critical for open source developers well being. Their public upstream work needs to be part of the internal sprint backlogs. Sprint retrospectives need to review the open source work along side the internal private project work. I have expanded on what is necessary here

Lastly, the most important part of your open source effort are your developers. They will be the beneficiaries the Open Source Office support. Your upstream open source developers will be productive contributing members of your commercial enterprise. Assign open source developers to public projects no differently than to private downstream projects. Do not forget the critical  part of a developer’s well being is good, clear leadership. You will need to have those developers at least dotted line to the Open Source Office leadership. This ensures that come annual reviews, your open source developer will get the proper attention and rewards for their hard work.

Remember that part-time work on multiple projects is possible for someone very senior.  Expect that junior to mid-level developers will need to work on one internal or external project at a time. I have many more details on upstream open source developer productivity here and here

The Open Source Office

In conclusion, the Open Source Office establishes the necessary engineering support structure for a commercial company to be successful. The Open Source Office mitigates risk and supports the developers. The Open Source Office responsibilities can be further described as legal and security risk mitigation, supporting open source governance, projects, and their developers. Stop the tug of war for engineering resources within your company and formally recognize your use open source software, by establishing an Open Source Office.

How to use Public Projects to Build Products

Lanai Rock with glass

Project to Product

I want to focus on one specific aspect of project to product development from the Open Source First article. Coordinating upstream (publicly licensed) and downstream (private non-licensed) development work. In any organization that is developing and supporting software, there are some number of engineers that work on public projects. It is a common problem that the Development Managers and the Scrum Masters do a poor job of tracking that public work. Let’s assume that the Development Manager is aware and supports the public development work.

Support the Developer

What typically happens is that upper management is not aware of what the upstream work purpose is and its importance to supporting the overall development strategy. What likely started as support or even direction from the Development Manager to work on the public project evolves into a fog of light understanding.

Many months later, when performance reviews and head count updates come around, typically the upstream developer gets lets left out, as the public work is not part of what upper management understands. This leads to either dropping the upstream work, alienating one of your likely best developers, and/or the upstream work continues, but with zero Development Manager insight. In some cases all three things happen, in which case, you now have a increasingly separated, forked development effort, with a increasing annoyed developer, that is going to be putting in more and more time into the public projects. Not a good workplace situation for the developer or the manager.

Treat Public Projects Like an Company

It takes more work, but there is a proven solution. Treat the public projects like another engineering organization. If you are going to work with another engineering team, you expect to have a clear understanding of responsibilities and timelines. You need to have the exact same expections when working with a public project.

Right at the beginning, when the developer has plans to work upstream, set out responsibilities and timelines. This is just as much to protect the developer as is it is the company. Define what the public work deliverable will be and when it will be delivered. Make sure your Product Managers understand how the public work contributes to delivering the product. Make sure the Scrum Master works with the upstream developer just like the rest of the downstream developers. The upstream developer is still part of the team. As the upstream work starts to take shape, get the code into a testable branch on your software pipeline. Make sure you have quality unit and build tests to verify the upstream work, so it can be more easily merged into your code base trunk when the time comes.

Track all your development work and report on progress. Highlight the upstream and downstream work. Update leadership on the upstream progress. Keep up to date with the public projects release schedule and strategy. Publish both your private release schedule along side the public projects release schedule.

Make It Happen

This is just the broad strokes. You will need to take this and put in specifics for your implementation. Organization structure and tools vary greatly, but hold to the basic tenants outlined above.

Public Product Management Transforms Private at Walmart

For those of you that know me,

know that I am more a of engineer than a product guy. Even so, I was hired on to transform the Walmart Platform Products using my engineering and open source experience as the Walmart Platform Product Director. It’s been very interesting so far. Not without some bumps and bruises along the way. I am happy to say that we are in our 5th month running an internal Platform Product Guild, largely based on the success we the OpenStack community have had with the OpenStack Product working group. We have development teams publishing Product Roadmaps that we regularly review across product teams and with customers. We are directing customer and engineering feature requests into our Product Roadmaps and product development cycles. We still a lot of quality to improve on, but we have made great progress so far. I am looking forward into 2017 to include more open source behaviors in the product development processes at Walmart. Stay tuned for more details as I can share them.

Why do we fall

OpenStack continues to have a lot of potential, because of the great work of the many technical and governance teams do release after release. However, there continues to be the appearance of releases not meeting operators needs. The lack of technical strategy is partly to blame here. Defcore, Product, Stable Release, Cross-project, and the Technical Committee fill gaps, but without a strategy, it’s the orchestra without a conductor. Well played music stepping on each other.

As evidence for the case of the missing technical strategy, I draw your attention to the OpenStack mission objectives of massively scalable and easy to install as long overdue objectives. Without the focus a strategy provides, these very public objectives have not been well met or even well discussed. We have put much attention on Defcore and protecting the OpenStack brand. It is very important without a doubt. I submit that brand protection through interoperability is an important, second priority. Interoperability makes the ecosystem of companies stronger, but does not fix the underlying problem that Defcore and other teams keep running into.

The Product team is well placed to take a strong position on what the technical strategy should be. The board can act on it by focusing members and directing the foundation staff. It would be a good objective for the Product team mid-cycle.

We need to recognize we have fallen, so we can get back up again.

OpenStack Vancouver Summit Talks, with EMC Federation speakers

Find the EMC Federation summit talks as is with current information below. Respond with comments and/or to the speaker directly. There will be a new post after the summit talk voting starts.

  Speaker(s) Session Type
(Talk, Demo, Panel)
Track Title Description (once submitted link to right, treat below as read only) URL for Submission
1 Shamail Tahir & John Griffith Talk Cloud Storage Cinder: Efforts in Cinder to provide quality as well as compatibility
Proposed Track: Storage
This session will provide an overview of the quality control and assurance procedures being taken in Cinder, specifically through third party CI. We will explain how the CI systems work, and what they test. We’ll also talk about why this benefits the end users and operators as well as discuss some of the challenges and learnings encountered along the way. This presentation will not be focused on a single Vendor or Driver, but is meant to discuss various drivers including the reference LVM driver.Key Take-aways for Attendees:
Understand the development workflow for Cinder (with an emphasis on quality)
Understand how Cinder tests and ensure interoperability with a large number of storage providers
Understand the emphasis on compatibility requirements for heterogeneous storage environments in OpenStack/Cinder
2 Shamail Tahir & John Griffith Talk Cloud Storage Storage Use-Cases in OpenStack This session will explain the multitude of storage options provided by OpenStack. We’ll talk about the differences between Object, Block and Shares as well as the options available to provide each of them. We’ll also explain the differences between persistent and ephemeral storage and how that relates to building and using Instances.Key Take-aways for Attendees:
Understand the various storage use-cases available in OpenStack
Highlight the purpose and high-level functionality of the OpenStack Storage Projects
Discuss the difference in storage consumption model from traditional IT
Gain an overview of how the existing OpenStack storage projects can serve your cloud needs
Walk-through a sample application architecture that leverages all storage services as a part of it’s stack.
3 Shamail Tahir & Sriram Subramanian & Gerd Prüßmann Talk Plan Your OpenStack Project Highly Available OpenStack: From Theory to Reality We all know there are two sides to every story, and, in this session the theory and reality of highly available OpenStack clouds will be discussed. We will cover concepts from the OpenStack HA-Guide that provide active/active and active/passive configuration guidelines (the “theoretical” configuration of OpenStack HA) and walk-through HA considerations that Deutsche Telekom had to implement, and design, from a practical production grade OpenStack deployment perspective (the reality). We will also discuss the state (and implications) of a multiple availability zone and region setup.Key Take-Aways:
Learn the tools and components you can leverage to make your cloud highly-available
Practical advice on implementation of HA OpenStack services
Identify resources available to help with the planning and design elements of your cloud HA strategy
Recommendations on selecting active/active or active/passive as a HA strategy
4 Randy Bias Talk Plan Your OpenStack Project State of the Stack v4 The original SOTS was the first end-to-end view of OpenStack as a project. It has over 90,000 views on slide share and growing every day. This is the fourth iteration of SOTS and we will cover the good, bad, and ugly of all integrated and core OpenStack projects. This session is ideal for anyone who is new to OpenStack, who desires to understand “the big picture”, or who is trying to get an honest “self-evaluation” of OpenStack as as a whole.
5 Erez Webman, Randy Bias Demo (we need someone who can put the demo together) Cloud Storage Highly scalable distributed block storage with ScaleIO Struggling with block storage for your OpenStack deployment? Over a third of OpenStack deployments use an open source distributed block storage solution that is difficult to deploy, maintain, has problematic performance characteristics, and can’t deliver on the promise of scale-out storage. EMC, the world’s leader in storage has a number of scale-out storage solutions, including ScaleIO, the product of over a decade of development by experience storage experts. ScaleIO is a 100% software solution that runs on commodity hardware. It is easy to deploy, maintain, and operate. It is highly performant and proven in the field at scale at over 100PB. ScaleIO’s built-in “protection domains” provide smaller fault domains, while ensuring performance across the entire cluster.In this session we will give a deep dive on the ScaleIO technology, and show it running on a cluster of 100+ servers doing massive IOPS and show how easy it is to manage faults by inducing several failures. We will also show how ScaleIO can be easily downloaded directly from the EMC website and evaluated for free. ScaleIO is plug and play with OpenStack and may solve some of the most egregious block storage problems you have had to date.
6 Sean Winn Talk Cloud Storage Software-defined Storage: the future of OpenStack storage systems This session will demonstrate how to use EMC ViPR as an option to build a multi-cloud SDS platform capable of handing heterogeneous storage environments. ViPR, when used in conjunction with Cinder, allows your storage platforms to be abstracted in a simplified manner while allowing storage operators to optimize resources through policy based management.
7 Jim Haselmaier, Niki Acosta, Shamail Tahir, Aaron Delp, Andre Bearfield Panel Enterprise IT Strategies Bridging The Gap Between App Developers, Operators & Technology: Product Management In The OpenStack Ecosystem As OpenStack grows incredibly quickly in popularity there is an increasing and consistent need to have a means to bridge the gap between the populations using and consuming OpenStack and the myriad of technologies and details that must be navigated when developing, producing and delivering OpenStack. In a mainstream business this function is filled by Product Management: Understand the needs of customers, work with Engineering to build the right product, and then communicate product information back to customers using terms they understand. The issue OpenStack faces today is that virtually any level of exposure to either the technology or the community almost immediately entails deciphering terms and concepts such as GitHub, repositories, trunk, open source, forks, CI testing, etc. etc. These terms are 2nd-nature to virtually all currently involved in the community, but are new and potentially intimidating to others. This is situation is producing unique challenges for three distinct groups:
The OpenStack Community: How do the people writing software for OpenStack best learn the needs of those that will be using it – when many of these prospective users do not wish to write OpenStack software themselves?
OpenStack Vendors: How do they manage the technology coming out of the OpenStack community and turn it into products that meet their customers’ needs?
Organizations Using OpenStack: With OpenStack’s new technology and cloud model (which often requires enterprises to behave in new ways) how do IT organizations change how they relate to and serve their user base so they can receive full benefit of this new environment?This panel discussion will explore these dynamics and provide practical perspectives on how these user/technology gaps can be addressed.
8 Sean Winn, Pradyumna Sampath (Hitachi) , Travis Wichert Talk Operations Have your Cake and Eat it too! Monitoring OpenStack in Layers One of the most important aspects for many cloud service providers is the ability to measure and predict cloud resource health and consumption. Withthe adoption of OpenStack on the rise in the Enterprise, it’s becoming more and more importantto be able to seamlessly integrate with existing tools and products that exist in the marketplace.As an operator, how will you be able to predict scale-out requirements? How can youensure availability and reliability? How can youdemonstratechargeback orshowback to your various groups who make use of your infrastructure? How do you budget and plan foradditional capacity? How does all of this integrate with your existing tools and skill sets?This talk will focus on the various integration points for monitoring your OpenStack infrastructure. We will be presenting a reference design which covers a layered approach to monitoring, allowing for operation at scale. We will also discuss opportunities where the community can contribute to the advancement of monitoring in OpenStack. You should leave this session with a better understanding of how to extract useful monitoring and telemetry data from your OpenStack deployment in order to operate at scale both efficiently and reliably.
9 Sean Winn, Mike Cohen, Naren Narendra, Sean Roberts, Allison Randal, Rob Hirchfeld, Stefano Maffulli, Arkady Panel Community What’s Next in OpenStack: A Glimpse at the Roadmap Whether you are a newbie to OpenStack looking at building your first cloud or an experienced operator with years of OpenStack success behind you, you’ve probably spent some time wondering what to expect from the OpenStack project over the next several releases. Will it finally support that new capability you’ve been waiting for? Should you plan for an upgrade in the next 6 months?While the development community is always working and planning new features, its takes a lot of time on IRC to get a complete view across the different projects. The OpenStack Product WG spent time this cycle working with the project teams and PTLs to understand their priorities for the next several OpenStack releases. In this session, we’ll present our findings across the different projects in an effort to give users a glimpse into the OpenStack roadmap.
10 Tushar Kalra Talk Things you need to know before you containerize your OpenStack deployment For those starting out, there is a tendency to treat containers like glorified light-weight VMs. When architecting your cloud around docker, you need to ensure that you adhere to some best practices to really leverage the benefits. This talk will go over some of those best practices and touch upon some of the challenges that you might encounter when deploying OpenStack services in docker containers.
11 Sean Roberts, Allison Randal, Rob Hirschfeld, Stefano Maffulli Talk State of OpenStack Product Management OpenStack is a very active community. Bursts of change happenquite often and itcan be difficult to keep up if you are not immersed. We need manage the flow of critical information anddecision making just like any other engineering organization. Thepeople that represent the Product Management of OpenStack are a critical group within the community.The Product Management working group has met a few times starting in Paris. This group has organized itself around three first activities. Gathering the current state of the OpenStack projects, Defining what the Roadmap could look like, and working with the Cross Project team. The User Stories from the Win the Enterprise working group will usedalongside the project needs.Join us to discuss what we have so far and let debate where we the OpenStack community should be going towards.
12 Sean roberts, Rob Hirschfeld, Egle Sigler, Alan Clark Panel DefCore, tempest DefCore 2015 This Committee was formed during the OpenStack Ice House Summit in Hong Kong by Board Resolution on 11/4. DefCore sets base requirements by defining 1) capabilities, 2) code and 3) must-pass tests for all OpenStack products. This definition uses community resources and involvement to drive interoperability by creating the minimum standards for products labeled “OpenStack.” Our mission is to define “OpenStack Core” as chartered by the by-laws and guided by Governance/CoreDefinition What has DefCore done so far? Who is involved with DefCore? What changes are planned around OpenStack branding in 2015? How will DefCore change OpenStack in general?
13 Sean Roberts, rest of ambassadors Panel Community Meet the Ambassadors oin to our panel talk about community status report and meet with the OpenStack Ambassadors. They connect the user groups to the Foundation, and help initialize the groups and guide them to grow.
Review of some Ambassadors launched actions during the last release cycle:
– OpenStack community report
– What is the size of the community ?
– Global and regional trends
– Introduce new groups, leaders
– Official group process
– Officials groups
– Process
– Examples of User group help
– Groups portal
– Overview
– Results
– Welcome pack and OpenStack shop
– Q&A
14 Sean Roberts, Stefano Maffulli, Loic Dachary Talk Community Community OpenStack Training Wants to Come a User Group Near You! We want to energizean OpenStack user group near you!To do that, we will be using the community training project. It ishappenning at the Tokyo and San Francisco user groups, Coming off the Paris summit, the community training guides project is focusing on the user groups as their main audience. While thereare paid OpenStack training programs available, the OpenStack Training Guides project aims to teach the basics of OpenStack through the user groups.In this talk, we will describe the Training Guides project objectives of OpenStack training cluster, training content mostly by HTML slides, example scenarios and use cases, and quizzes. We will walk through our successes to date delivering training thorugh the user groups.
15 Tyler Britten / Mohamed Khalid Talk Cloud Storage OpenStack Data Storage Deep Dive With the almost limitless storage configuration options in OpenStack, architecting, operating, and troubleshooting can be daunting. In this session, we’ll cover configuration best practices, operational tips, and troubleshooting techniques with real-world examples. We’ll also discuss the various storage projects in OpenStack- Cinder, Swift, and Manila, and how EMC is contributing to them as well as how we are integrating our storage products.
16 Tyler Britten / Mohamed Khalid Talk Operations OpenStack Operations While there is a lot of information available about the various OpenStack deployment options, there’s surprisingly less about what to do once you have your OpenStack environment up and running. In this session we’ll talk about all of the hot-button OpenStack operational issues- high availability, upgrades, monitoring, troubleshooting, and more!
17 Drew Smith Talk Related OSS Projects Architecting OpenStack Monitoring for Next-Generation Functionality Monitoring and alerting: two things that everyone, operator to CTO, can agree are critical parts of any production deployment. In this presentation we’ll discuss the different generations of monitoring technologies – where we’ve been and where we’re going – and give a high-level overview of the current efforts and difficulties within the OpenStack ecosystem. We’ll talk about the importance of the shift away from polled service checks towards ‘push metrics’ and active telemetry, and present some concept designs for some seriously cool operator / administrator features that will be made possible in the near future.
18 Patrick Butler Monterde, Adrian Moreno, Magdy Salem Talk Related OSS Helios Burn: A REST API Fault Injection Platform Helios Burn is an out-of-the-box REST fault injection platform that captures and modifies HTTP/S traffic. It implements a man-in-the-middle interception using self-signed certificates to be able to intercept and interpret HTTPS traffic.The purpose of Helios Burn is to provide developers with a tool that injects failures in REST APIs so that developers can verify the stability and resilience of their applications and identify and prevent failures before deploying them into a production environment.HeliosBurn let’s users create custom rules to match with the REST target traffic, or they can also benefit from the preset rules for common Cloud services including OpenStack Swift and Nova. Upon a match, users are able to apply actions such as modifying any HTTP information (i.e., headers, URL, status code, payload), respond on behalf of the server, delay the request or response, or drop the connection.It is designed with an extendable modular architecture that enables third parties developers to add new modules with custom functionality.HeliosBurn is managed through a friendly web dashboard that allows users to tweak any aspect of the platform and observe the HTTP traffic going back and forth. In addition, HeliosBurn provides a full-featured API for developers to create custom clients and libraries.HeliosBurn is shipped both as a VM and a Docker microservice, making it really easy to deploy it. Depending on the need, it can be placed in a standalone server, in a Virtual Machine, or co-located with the Web server or Client application.As an open-source project, HeliosBurn welcomes and encourages any kind of collaboration from the community.
19 Shamail Tahir, Dan Wendlandt Talk Enterprise IT Strategies Leveraging Your Existing DC Investments Most enteprise customers are transitioning or augmenting their IT strategies with OpenStack. In this session, we’ll discuss how to repurpose, or leverage, your existing IT investments in your OpenStack project. We will discuss how workloads may influence which assets to leverage, how to “pilot” OpenStack”, and start using OpenStack with the minimal amount of net new investment.
20 Marcos Talk Networking Leveraging vSphere Virtual Distributed Switches and NSX vSphere for Neutron In this session you will learn about the Neutron plugin that VMware has developed and released to the community. This plugin allows supports basic and advaned Neutron workflows and leverages the NSX vSphere solution for added flexibility and scalability in your OpenStack Cloud. Logical Switching, Logical Routing, Distributed Firewalling are all NSX services that can be consumed by Neutron and exposed to your cloud tenants.
21 Marcos HOL Hands on Lab Guided Lab for Learning All Aspects of OpenStack Do you want to learn and use OpenStack APIs? Do you just want to get hands-on experience of using Heat templates or Neutron Networking? Want to learn how OpenStack integrates and runs on VMware technologies such as vSphere and NSX? Is there an architecture that I can check out to see what all components are need to run OpenStack in production (Message Queues, Memcache, DBs, Load Balancer…etc)? Curious how you would monitor and troubleshooting your OpenStack Deployment? Merge this with Hands on Lab for broader OpenStack + VMware. Whether you are curious to learn about OpenStack or how it works on VMware. This hosted lab gives you the perfect oppoertunity to learn all aspects of OpenStack.
22 Dan W Talk IT strategies Unicorn Stack
23 Dan W Talk IT strategies OpenStack for VMware Admins As OpenStack continues to grow, Enterprises are beginning to explore and to implement OpenStack as their Cloud platform of choice. Often, these companies have existing investments and expertise with VMware technologies. In order to prepare for this new world, these people who are familiar with VMware concepts and terminology will need to understand the parallel concepts and terminology in OpenStack.This session will be valuable for anyone who needs a better grasp of how to talk about both VMware and OpenStack in an enterprise context.
24 Eric Lopez, Aaron Rosen, Janet Yu HOL Hands On Lab Openstack Networking Introduction Hands on Lab This session is an introduction to new users on Openstack Networking. Users will be provided access to a live Openstack environment with Neutron setup. We will walk through
the key neutron deployment use cases with members of the Neutron core development team available to provide guidence and answer questions.Demonstrated features will include:
– Creation of tenant networks using overlay tunnels.
– Configuration of external connectivity
– Advanced Neutron Features, including support for overlapping IPs,L3 + NAT usage via logical routers, Firewall as a Service,Loadbalancer as a Service, VPN as a Service, IPv6 and more!We will incorporating lessons learned from presentation of this session at previous Openstack Summits and also including new Neutron capabilities introduced in the Kilo release.
25 Eric Lopez, Aaron Rosen, Janet Yu HOL Hands On Lab Openstack Networking Advanced Hands on Lab This session is an introduction for operators on Openstack Networking. Users will be provided access to a live Openstack environment to install and configure Openstack Networking Neutron.
We will walk through configuration of Neutron with the ML2 plugin via OpenvSwitch(OVS) and L3 services with OpenvSwitch Virtual Networking(OVN).Demonstrated features will include:
– Interaction with other OpenStack components (Compute & Storage)
– Configuration of Metadata Services and DHCP Services
– Designing Neutron for HA
– Troubleshooting NeutronThis session highlights how the environment is configured for Openstack Networking Hands on Lab at previous Openstack Summits.
26 Eric Lopez, Aaron Rosen HOL Hands On Lab Congress Congress is an OpenStack project that provides policy as a service across any collection of cloud services in order to offer governance and compliance for dynamic infrastructures. In this lab users will get access to a live OpenStack setup with congress already installed and will be able to walk through several key congress deployment use cases and get hands on experience working with congress. Users will write policies that interface with several OpenStack projects (neutron, glance, nova, keystone, cinder, murano) and understand how the policy language works and how one can tame their cloud with congress.
27 Talk Community Lessons from the San Francisco OpenStack User Group
28 Tim Hinrichs Talk Related OSS Projects State of Congress
29 Jay Jahns Talk Operations Practical Lessons from real world Multi-Hypervisor deployments
30 Ryan Hsu Talk How to Contribute How to run and upkeep a 3rd party Openstack community CI
31 Tim Hinrichs, Serg Melikyan Talk Operations Governing (Murano) Application Deployment with (Congress) Policy Deploying applications is hard to get right. It requires gathering information from many different resources (e.g. the application itself, the infrastructure, the other applications already deployed), and making technical and business decisions about where and how to deploy the app while satisfying the multitude of business/infrastructure/application policies that govern the deployment process.In this talk, we describe an integration of Murano and Congress that eases the burden of policy-governed application deployment. This integration ensures that application-deployment done through Murano complies with the policy expressed in Congress—from initialization all the way through to final deployment. In this session we demonstrate how to define policy with Congress and how policy is enforced within Murano during application fulfillment, culminating in a live demo.
32 Tim Hinrichs, Ramki Krishnan Talk Telco Strategies Helping Telcos go Green and save OpEx via Policy Currently OpenStack does little to help Telcos optimize their workloads for energy consumption, cost, and speed. Today, operators must manually (or via scripts) provision, migrate, and decommission workloads to achieve the desired balance of energy/cost/speed, and they must do so repeatedly.In this talk, we describe an open architecture for automating resource optimization, where operators provide a policy describing how workloads ought to be optimized, and OpenStack continually monitors and migrates workloads to satisfy that policy. Under this architecture, operators give their policy to Congress [1] (the not-yet-incubated OpenStack project for Policy-as-a-Service), and Congress continually enforces that policy by migrating workloads as appropriate. In addition to discussing the architecture, we demo a proof-of-concept implementation where Congress migrates real VMs via Nova in response to changes in datacenter readings reported by Ceilometer.
33 Dimitri Stiliadis Tim Hinrichs, Mike Cohen, Dave Lenrow, Craig Matsumoto Panel Related OSS Projects Panel: Defining Policy Frameworks for Openstack Policy has quickly become a hot topic in cloud management and orchestration. As OpenStack clouds expand, penetrate the enterprise, and evolve with technologies such as containers, policy-based solutions for capturing user intent, automating management and security, and ensuring governance and compliance for applications has emerged as a critical area for development.
This panel will explore emerging trends and projects in policy developing in the OpenStack and OpenDaylight communities. It will discuss a number of topics, including:
-What is meant “policy” in the context of OpenStack. Is there a “right” approach?
-Why is policy important? What are the key use cases?
-What projects and capabilities are present in OpenStack today
-How will it fit with existing OpenStack components
34 Somik Behera Panel Networking User Panel: Neutron Considerations in Production environments
35 Somik Behera, Gurucharan Shetty Talk Networking Container Networking models with OpenStack Neutron
36 Somik Behera Talk Networking Neutron – Past, Present & Future of Cloud Networking
37 Boden Talk Products Tools Services Choices of deploying OpenStack on VMware VMware is now serving a plentiful menu of OpenStack delicacies; bound to satisfy the appetite ofa wide range of customers. Whether you fancy a small-to-mid sized prescriptive OpenStack deployment drizzled atop your existing VMware based technologies in a matter of minutes, or you have an intense craving for a highly customized large OpenStack deployment; VMware’s got you covered.In this meal we’ll sample VMware’s OpenStack menu which ranges from a click-and-go out-of-the-box integrated OpenStack distribution, to a highly customized made-to-order OpenStack masterpiece. We’ll dive into the ingredients of these recipes to better understand VMware’s common OpenStack reference architecture, how the solutionis deployed / operated and how VMware offers a OpenStack based Software Defined Data Center (SDDC) solution for Cloud appetites of all sizes. For dessert we’ll indulge with some details on the custom integration between OpenStack and VMware technologies; making your OpenStack meal service a pleasant and affordable experience.By the end of this feast you should walk out fully satisfied with an understanding of how VMware cooks a delicious dish of OpenStack to suit any occasion you may have.
38 Somik/Aaron/Bruce D? Talk Telco Strategies Considerations for Containerized VNFs with Docker & OpenStack
39 Ryan Hsu Talk How to Contribute running a remote CI infra (Ryan)
40 Somik Talk Products Tools Services Production Neutron deployment Architectures with NSX (Somik/NSX SA/SE/Dimitri)
41 Somik Talk Products Tools Services Advanced Networking Deep dive with VIO ( NSBU TPM)
42 Justin Griffin, Dan F Talk Products Tools Services heterogenous vROPS, Log Insight
43 Somik Talk Related OSS Projects Open Virtual Network (OVN) for OpenStack (Justin P/Ben P.)
44 Somik Talk Cloud Security Implementing zero-trust micro-segmentation architecture with Neutron ( Srini N. & Shadab Shah)
45 Tracy Jones Talk Compute Diving Deeper into Nova and VMware ESXi It is well known that Nova works with the VMware hypervisor. Yet, there is quite some confusion around how Nova integrates with VMware ESXi. Does Nova interact directly with ESXi or with the vCenter Server? Is Nova capability X supported when using the VMware hypervisor? Can I take advantage of ESXi/vCenter Server’s feature Y from Nova? When should an admin use Horizon and when should one use the vCenter client? This talk will mitigate such confusion by digging into the nuts and bolts of the integration between Nova and the VMware hypervisor with the help of a demo that will also show case how some of the advanced ESXi/vCenter Server features can be leveraged from Nova.
46 Dan F Talk Storage vSAN (Nexenta?): VSAN for Cinder & Nexenta for Manila/object storage
47 Dan F Talk Storage VIO + SwiftStack
48 Dan W Talk User Stories Adobe (Frans plans to submit)
49 Sean Roberts, Sharmail Tahir, Tim Hinrichs Talk Planning your OpenStack Project Leveraging Congress for Policy Management In this session, we will cover how Congress can be leveraged by your organization to ensure compliance and policy adherence in your OpenStack cloud. The example governance scenario will show how to set and monitor policies for compute, network, and storage.
50 Sean Winn, Ted Streete Talk Networking Building Scalable Networks for Fun and Profit So you’re going to use Neutron plus an SDN overlay? Or perhaps just simple VLANs? Regardless of which way you go, it turns out there is already a set of well understood best practices for building scalable networks. In this session folks who have built scalable networking for large OpenStack deployments will walk you through the dos and don’ts of networking. Why layer-3 networking is your friend, how OSPF and BGP work together, and why everyone loves a spine/leaf networking architecture. We’ll give real world examples of networks we have built, including one that handled the load for a major retailer during Black Friday 2014.
51 Aaron, Tim, Alex, eric Hands on lab Congress hands on tab Congress is an OpenStack project that provides policy as a service across any collection of cloud services in order to offer governance and compliance for dynamic infrastructures. In this lab users will get access to a live OpenStack setup with congress already installed and will be able to walk through several key congress deployment use cases and get hands on experience working with congress. Users will write policies that interface with several OpenStack projects (neutron, glance, nova, keystone, cinder, murano) and understand how the policy language works and how one can tame their cloud with congress.
52 Aaron, Pierr, Ramiro Salas, Pierre Ettori; Enterprise Extending OpenStack Congress to the PaaS layer for Next-Gen App Policy Controls Cloud Foundry is a open source cloud computing platform as a service (PaaS) which allows users the ability to deploy and scale their application easily. The platform provides several as as a service features such as redis, mysql, mongo, etc that application developers can leverage with their applications. Congress integrates with Cloud Foundry and allows Security teams to audit and define policies over specific applications. For example, a security team could define a policy saying applications deployed in production require https or a specific autoscaling configuration. In this talk we’ll give an overview of congress and demo this integration.
53 Randy Bias, Sean Roberts Talk Community Building DefCore and Me DefCore is a recently formed set of criteria that identifies which products, providers, and solutions meet the requirements to use the OpenStack mark. In this session, we will explore the considerations and implications from one vendors perspective as they begin to assess their own readiness under this new program. We will also discuss why OpenStack cloud operators can also benefit from validating their implementation against DefCore using a tool called RefStack and how this initiative will help the compatibility of OpenStack clouds in the long-term.