Mirantis Spring OpenStack Summit Proposals, Part 3

On Friday we brought you part 1, a selection of proposals Mirantis is putting forward for the Spring OpenStack Summit in Atlanta, along with a complete list of titles, and Monday we gave you part 2.  Today we wanted to post the rest of the list.

Here's today's selection:

• OpenStack LBaaS ecosystem: coopetition in work (Samuel Bercovici,Eugene Nikanorov) - Before the Havana release, LBaaS was limited to open source HA proxy implementation only. The Havana release also opened LBaaS  to commercial and enterprise grade load balancers  via a driver mechanism, that basically creates a standard API consumed by the different load balancing vendors' LBaaS drivers.  Maintaining  API  standardization is critical for creating healthy ecosystem. In this session we share the process of collaboration behind the definition  and implementation of L7 content switching and SSL termination as standard LBaaS APIs. (Read More and Vote)

• Software Defined Networking Performance and Architecture Evaluation (Brian Chong) - Over the last 3 months, Symantec has run extensive performance and diagnostics tests across multiple overlay providers as well as against the base Neutron VLAN configurations and have come to several insights into CPU penalties, Network design issues at scale as well as performance comparisons using different encapsulation techniques. We will explain what our architectural design was, the results of our testing as well as our design insights into how Symantec’s cloud will be affected by the outcome of the evaluation. This presentation will cover architectural and performance characteristics of what a Overlay network is compared to a straight VLAN based environment on a cluster size of 100+ nodes. (Read More and Vote)

• Is your OpenStack Cloud Compliant? (Evgeniya Shumaker) - Compliance with critical industry and regulatory standards (like PCI DSS for payment card security, HIPAA for healthcare information, SOX for corporate process management and auditability, etc.) used to be mostly the concern of industry-specific application makers and customers integrating their solutions. Cloud computing – especially IaaS – has made things a lot more complicated. As clouds supplant conventional physical infrastructure, private cloud builders and cloud service providers are having to shoulder the burden of maintaining standards-compliance around the workloads they host and the organizations they serve. How do you architect and build standards-compliant OpenStack clouds? This presentation offers an inside look at the process: the most important standards for cloud builders,where existing OpenStack resources can fully or partially solve common compliance problems,where standards support within OpenStack is currently thin,the workflow for architecting standards-compliant clouds and common risks and emerging opportunities. (Read More and Vote)

• KeyStone Security and Architecture Review (Brian Chong) - In this presentation, KeyStone Security, specifically Trusts or Delegations, AMQP Security with KeyStone and integration with a Corporate LDAP will be discussed. The nature of OpenStack KeyStone plays a major role in binding all of the Projects together but not much is mentioned about how to do this with KeyStone or what the pitfalls and dangers of hooking up a centralized Security System to the rest of the cloud will be. The security and protection of the Identity and Token repository for OpenStack must be a well protected component within your Cloud Infrastructure. After this discussion, you will come out knowing more about KeyStone Security and how you can use it to your advantage. (Read More and Vote)

• Python and Performance: What and How (Dmitriy Ukhlov,Serge Kovaleff)   How much you can do to optimize code inside a single node? With the Python WSGI app you can use horizontal scaling to your advantage. In this talk we share our experience in the optimization of MagnetoDB, which provides high-performance Key/Value storage as a service, and what we learned about how much you can do to optimize code inside a single node. WSGI framework research and comparison,uWSGI and Gunicorn in multi-core computing environment,tuning Gunicorn and GEvent patches for Cassandra usage will all be covered in this discussion. By the end of this talk, you will know how to get the most out of your Python WSGI app. (Read More and Vote)

• Advanced services in OpenStack Networking (Eugene Nikanorov) - The presentation will cover usage patterns of advanced networking services as well as some core functionality that develops into a multivendor environment. In this discussion, we will talk about the following topics: load balancing, firewalling, vpn a l3 service and modular l2. The primary focus of this presentation will be on multivendor usage. (Read More and Vote)

• Extending the Fuel project to accelerate and simplify deployment of Openstack drivers and components (David Easter) - Fuel solves the problem of installing and deploying OpenStack by providing device auto-discovery, pre-deployment validation, automated cross-network configuration and deployment, and cluster sanity and functional testing -- all under a web-based GUI that simplifies decision-making, prevents certain kinds of errors, and helps the user visualize critical information. Fuel lives upstream from OpenStack and currently comprehends multiple OpenStack releases ‘out of the box’, but more importantly, Fuel is open source. How does one go about contributing to Fuel? In this session, we’ll fully discuss the following with attendees: the structure of Fuel, controlling the Fuel UI, how user actions control what Fuel does, how to create Fuel wizards, how to support additional OpenStack features and how to control what Fuel does behind the scenes. (Read More and Vote)

• Best Practices for Ceph Deployment with OpenStack (Dmitriy Novakovskiy, Dmitry Borodaenko) - Ceph has long promised to provide a platform to satisfy all OpenStack storage needs, and last year, it delivered.  Now we want to share the good, the bad, and the ugly sides of what we've learned. In this talk, we will explore the reasons to use Ceph, go through things you should consider in your storage architecture, tell you how to avoid the most common deployment and integration problems, and describe the reliability and performance best practices that will help you get the most out of your Ceph backed OpenStack environment. (Read More and Vote)

• OpenStack Distributions: Do We Really Need Them?  (Dmitriy Novakovskiy) Lots of people build private clouds with OpenStack these days, and more will begin in the near future. “Where will you consume OpenStack from?” It’s only been a few years, and a number of options are already available, you can consume stable release from the Community Trunk, you can rely on a vendor to provide you with an OpenStack distro or you can wait for the Community to produce “vanilla” packages/images and deployment framework.  The choice is not obvious, and always depends on your company’s context, IT capabilities, and business priorities. In this summit talk, we’ll discuss the key ideas and trade-offs of each option, and look at the risks and things you need to keep track of after a decision is made. Once you've attended this talk, you’ll have a good basis for advising your company on where your OpenStack should actually come from. (Read More and Vote)

• Performance of Hadoop on OpenStack (Andrew Lazarev) - Savanna provides integration and automation for Hadoop deployment on OpenStack, but what about performance in this environment? In this talk, we will discuss the performance impact of running Hadoop in a virtualized environment inside an OpenStack cloud. What is the performance difference between bare-metal and virtualized Hadoop environments? What is the IO overhead in OpenStack? How does data locality influence data flow? This presentation will cover these and other questions about the performance of Hadoop on OpenStack by presenting a series of benchmark tests over different installations of Hadoop. (Read More and Vote)

• Neutron NameSpaces and IPtables - Icehouse Update (Damian Igbe) - In the Grizzly release, using the OVS plug-in with Quantum required creating an additional Linux bridge for every instance spawned.  Not only was this messy to manage, but it also increased the number of devices a packet must pass through when making the trip from the originating instance to the Internet, and thus network overhead also increased. The Havana release has an improved approach, eliminating the need to propagate an extra Linux bridge per instance, since the OVS plug-in can now apply IPtables rules on Nova instance virtual NICs (i.e., tap devices). We will explore how the Havana implementation affects the underlying IPtables rules and security groups using the ML2 Neutron plug-in. We will build on material presented at the Icehouse summit by delving into how OpenStack Neutron implements security using IPtables. We'll also drill into namespaces. Security groups are built using IPtables.  When you attend this talk, you will have a good understanding of IPtables, how they are used to implement security groups, and namespaces in Neutron. (Read More and Vote)

• Python OpenStack Clients for cloud administrators (Christian Huebner) - In addition to command line tools and REST API calls, Python OpenStack Clients can be used to examine and control clusters. This collection of clients gives developers and devops a clean interface for developing Python applications to build, maintain and troubleshoot OpenStack clusters. In a recent project we developed an application that needs to acquire cluster configuration across all major projects. We found that Python OpenStack Clients provided a very useful interface to the project's cluster. This talk aims to provide you with a starting point to develop your own applications and an overview of the pitfalls we encountered on the way. To get the best results from this talk you should have basic Python skills and some OpenStack cluster administration experience. We will show practical examples for Keystone, Nova, Glance and Neutron on a demo cluster that can be replicated by you and used as a base for developing your own tools. (Read More and Vote)

• Climate, a Resource Reservation Service for Openstack (Sylvain Bauza, Dina Belova) - OpenStack is growing in multiple directions, with new requirements arising almost every day. For the use case of a large internal private cloud shared among a wide range of users, we definitely need to think about reserving resources. Climate is born from the idea of being able to reserve any type of OpenStack resources, either now or in the far future, in a predictable way, and we really believe that this project will help make OpenStack clouds more flexible and manageable. This Stackforge project, which aims to be incubated in the next couple of months, currently provides two distinct types of resources: virtual instances and compute hosts. In this presentation we’ll go through different reservation concepts provided by this leasing service, the current features, and the roadmap for future phases of the project. Also, we’ll present a demonstration of how Climate can help you with your resource reservation needs, so you’re welcome to join us. (Read More and Vote)

• Mistral: OpenStack Workflow Service (Dmitri Zimine, Renat Akhmerov) - Mistral is a task-automation system for OpenStack - “‘Workflow as a service.” OpenStack and application developers can use this powerful tool to coordinate complex distributed processes. Mistral will manage the entire automated process, adapting as necessary to insure the task completion.  Cloud administrators and operators can use Mistral as a flexible automation tool to express their automations as scalable, highly available workflows, managing task scheduling, cloud environment deployment, migration, or riding herds of long-running processes, like big data analysis and reporting. In this session, we share Mistral vision, use cases, and architecture, show the demo, and present current status and roadmap.  (Read More and Vote)

• Yet Another Query Language - Querying Complex Datasets in OpenStack and Beyond (Alexander Tivelkov) - The need for querying and manipulating data structures comes up regularly in most programming scenarios. It is even more common in OpenStack, as cloud infrastructures are usually huge, have many interconnected entities, and are organized in complex hierarchical structures with nested properties and relationships. While working on the Murano and Mistral projects for OpenStack, we realized the need for a flexible and extensible query language, able to run complex filtering-and-search tasks with a single and expressive line of code, so we built just such a tool and called it YAQL - Yet Another Query Language.  In this session, we will provide an overview of this language and demonstrate how to use it. At the end of this session, you will understand how to query a Python data structure anywhere in OpenStack using YAQL.  (Read More and Vote)

• Writing Ironic vendor extension (Maksym Lobur) - Ironic is an OpenStack project which provides an API for managing and provisioning physical machines. And the Ironic driver for Nova gives this power to Nova. So why should hardware vendors be interested in Ironic? The answer is that each feature provided by Ironic relies on hardware support. Some examples are: to manage baremetal node power, we use IPMI (Intelligent Platform Management Interface) and to deploy baremetal node we use PXE (Preboot eXecution Environment). The trick is that Ironic does not force you to use these interfaces, they’re just default. If you are selling hardware that provides better options - perfect! You can utilize your advanced hardware features inside your Ironic vendor driver and gain one more pro to convince your customer. After you attend this talk, you’ll know how to provide your hardware features for the future self-managing OpenStack infrastructure, in particular how to write your own Ironic vendor driver. (Read More and Vote)

• OpenStack on White Box Hardware (Russ Lindsay) - Frictionless access to information makes everyone think they can do this at home. But, in fact, deploying OpenStack on white box hardware presents a handful of challenges that your infrastructure team might not have thought about. In this session, we'll look at a range of challenges that can stop your project before it makes it to deployment. The session is divided into two topics that we’ll dive into: introduction to scalability techniques:clouds from a hardware perspective and the second section of the session will focus on strategies to simplify the white box experience.  People who will attend will also learn about some of the simple things that are obvious with enterprise hardware but can be challenging in any white box environment. Examples include power on/off, power monitoring, system health/event monitoring, and remote console access. You'll leave this session armed with practical knowledge about running OpenStack on Open Compute and other white box hardware. (Read More and Vote)