OpenStack needs you! Get voting for our ODS Vancouver talks!

We’ve submitted over 30 talks to the upcoming ODS in Vancouver and really need your help to make ensure they are included in the schedule. We’ve listed them all below with links to where to vote for each talk.

OpenStack is big and everybody wants to be a part of it. More and more vendor organisations want to ride the OpenStack wave and see getting talks on the summit schedule as a key way to strengthen their position as an OpenStack mover and shaker. The net result is 1500+ talk submissions which the current selection process cannot handle. There has to be a better way, join Mark Baker to discuss problems with the current process and how it might be improved.

Vote Now

For a long time big Enterprises have spent big buck on Enterprise software to support their enterprise activities. Smart organisations are starting to realise that their resources should be focused where business value is created – the areas where they can create a secret sauce that will give them an advantage over competitors. Every other aspect of their business should utilise commodity as much as possible to minimise cost and increase flow of resources in secret sauce creation. Implementing cloud gives your organisation an opportunity to examine where value is created and which technology services should be moved across to commodity apps and infrastructure, after all, if you take your existing stack and put it on cloud then you are failing to evolve, and we all know how that ends.

Vote Now

At every OpenStack Summit we hear complaints that various aspects of OpenStack are hard and complex: Where are the reference architectures? Where are the simple point and click installers? What are best practises for implementing HA? How can we work with a 6 month release cycle? In general there are good answers for most of these questions, but clearly we are not making the answers easy enough to find or the tools easy enough to use. This talk will show some of the proof points of how ease of use drives adoption some of the resources available that help make OpenStack easy and how we can continue to improve this process.

Vote Now

For a long time big Enterprises have spent big buck on Enterprise software to support their enterprise activities. Very few monolithic enterprise applications are appropriate and implementing cloud gives your organisation an opportunity to commodity apps and scale out infrastructure, after all, if you merely take your existing stack and put it on cloud then you are failing to evolve, and we all know how that ends.

Vote Now

At present, Container/LXC technology has becoming increasingly popular in the cloud computing industry due to the advantages of higher density and extremely faster bootstrap speed. As a sequence, more bigger virtual L2(Layer 2) will take away the ability to continue to use traditional broadcasting way. The separation idea of SDN is to connect overlay/openflow network intelligently between control plane and forwarding plane by disabling arp broadcast and learning ip/mac mapping automatically to largely enhance the performance of the whole datacenter network. This session will cover the following topics:

1. Container networking introduction
2. Network design options for container networking
3. Example setup of high density environment based on openstack & docker with L2 overlay and SDN technology
4. Performance evaluation and improvements

Vote Now

As the preferred platform for developing and deploying OpenStack, Ubuntu has had a strong focus on ensuring the quality of OpenStack releases on Ubuntu since its first inclusion in Ubuntu 11.04. With multiple Ubuntu and OpenStack release combinations to support, scaling the automated testing of Ubuntu OpenStack and the Juju OpenStack charms for deploying OpenStack has been a challenge the Ubuntu OpenStack team has met over the last few years. Come and hear how we’ve scaled our existing hardware testing capability by testing OpenStack on-top of OpenStack, including leveraging Neutron to provide L2 separation between testing environments, to ensure that we can support the multiple OpenStack/Ubuntu release combinations that support Ubuntu OpenStack.

Vote Now

OpenStack is a complex system with many moving parts; devstack has provided a solid foundation for developers to test OpenStack, and has been an essential part of the gating process since OpenStack’s inception. devstack typically presents a single-node OpenStack deployment from source – this has testing limitations as it lacks the complexities of a real multi-server OpenStack deployment. Ubuntu now addresses the complexity of multi-node service orchestration of OpenStack deployments and is growing support for deploying OpenStack from source rather than binary packages; by utilizing this feature and the support in Ubuntu for KVM and LXC containers, we can build a simulated multi-node deployment in a single physical machine, enabling developers to test their code in a more production like environment with similar ease of use to devstack. Come and hear about how we’ve implemented this feature for Ubuntu OpenStack development, how its used and see how you can instantly deploy code changes into a multi-node OpenStack deployment on your laptop!

Vote Now

Docker, Rocket, LXD, OpenVZ, Kubernetes, systemd-nspawn, lmctfy. There has been a lot of excitement recently about containers. Further, there have been a lot of announcements of container-based tools and toolchains, each with their own design goals. This is a rapidly changing space that can be difficult to follow and understand; at the same time, it is clearly very exciting and important.
In this talk I’ll try to define what people in the container community mean by a container, and position each of the container frameworks on a spectrum. I’ll give a basic introduction to namespaces, cgroups, and the various LSMs which make up a container. I’ll give an introduction to the idea of an app container and a system container, and talk about applications for each. As for the dummies part: I’ll do all of this using pictures of cats.

Vote Now

OpenStack is made up of a several services that interact with each other to provide a massively scalable cloud operating system. You likely already know the complexity that is involved with OpenStack deployments, particularly if you’ve had the opportunity to manually deploy OpenStack. This session will look at the conceptual architecture of OpenStack and show what we’re doing to simplify deployment, management, and scaling of OpenStack.

Vote Now

High availability of OpenStack services within a OpenStack Cloud is key for ensuring availability of cloud services to end users; The OpenStack Charms for deploying OpenStack on Ubuntu using Juju have now been through 18 months of battle testing in HA customer deployments – this presentation will detail the overall reference architecture of the OpenStack Charms in HA mode and how we have addressed the challenges of automatically deploying and scaling OpenStack services in highly available configurations.

Vote Now

OpenStack’s growing and maturing project and user bases demand thorough and continuous testing against reference architectures and production scenarios. This talk will be about how the Ubuntu Server Team automatically stands up and tests 100+ private OpenStack clouds per day in different configurations using open source service orchestration and machine provisioning technologies, both in bare metal and in nested OpenStack-on-OpenStack environments. Topics of deployment testing, hypervisor testing, trunk testing, regression testing, test framework testing, and deployment framework testing will be covered. And yes, there will be a test.

Vote Now

OpenStack is meant to work across a variety of different vendor platforms seamlessly — how do we ensure that it does? This talk will focus on wrangling the different OpenStack services (twiddling the various knobs) onto actual hardware. This covers choosing the services to deploy; dealing with physical hardware; automatically creating a tempest config from an existent cloud; and understanding over 100 tempest runs per day.

Vote Now

Many users are looking for ways to have IPv6-only Openstack deployments. This topic will be about highlighting the advantage of using ipv6 vs. ipv4 and what needs to be done to achieve robust IPv6 support in (at least) the main OpenStack modules i.e. Keystone, Glance, Nova, Cinder, Swift as well as common storage backends such as Ceph, other useful technologies such as haproxy, rabbitmq, pacemaker/corosync and so on. We will discuss the problems we have encountered and ways that we have found to overcome them as well as some suggested improvements that could be made to better support IPv6 across these technologies.

Vote Now

Nova supports a number of image backends (raw,qcow,lvm,rbd) but does not support direct attaching physical disks to Nova instances. The ability to colocate guests with physical storage tiers would enable instances to benefit from raw performance of local storage that network attached storage may not be capable of delivering. One proposed solution would simply leverage and extend existing placement support either between Cinder and Nova or simply within Nova itself.

Vote Now

OpenStack is a complex system with many moving parts. Thanks to the great strides the OpenStack community has made over the last few cycles, deployment is no longer overwhelming. But how do you manage that deployment over time? How do you add capacity and deal with change? Come and hear how we’re answering these questions with the Ubuntu OpenStack Autopilot. Choose your preferred components. Select your hardware resources. Click a button to roll it out. Then watch the Autopilot guide you through the management and scale-out of your deployment.

Vote Now

Learn from an experienced high availabilty expert (SAForum, Scope-alliance, PICMG, OpenSAF member) how “I want high availability” reflect in reality. Furthermore what it now means a Cloud based infrastructures. Debunk myths versus reality of what the industry needs, spoiler: it is not necessary what they asked. For the operators or corporation in needs try to understand the differents options to help get the opensource community develop (and test ? ) those features for you. From design to operation high availability is achieve by standing on an handfull set of principles. Aim of this talk is to facilitate collaboration between large corporations used to high availability for decades with the opensource culture, methods and behavior.

Vote Now

Former Unix adinistrators are evolving to Linux and OpenStack environments. New challenges needs to be faced and new tools are needed. Transition from Unix to Linux explained in an easy way. Old unix users are very welcome to this community

Vote Now

The OpenStack IT Lanscape has changed significantly in the last two years. New processor architectures (new and legacy) are included in OpenStack and the options available for the customers have increased a lot. Composite architectures are now possible: ARM + Intel + Power can be conbined to maximize the performance and reduce the cost.

Vote Now

In this talk we will show how to deploy OpenStack on a Power server on a simple and easy way, showing the differences with other platforms.

Vote Now

You’ve set up OpenStack and are ready to deploy workloads and really enjoy the elastibility the cloud affords you. Which instance sizes should you use? How should you configure your disks? Does your workload perform better with one set of networks, disks, or hardware architectures? Explore the world of repeatable benchmarking of workloads across instance types, regions, and even cloud setups.

Vote Now

As you plan for and execute your OpenStack installation it’s important to make sure you pick components that meet your needs for your workloads. In this session we’ll go over how to benchmark components in any permetation of OpenStack install in order to best select which component works for your hardware with your planned workloads.

Vote Now

How Canonical and Data Foundry improved time to marked for Golded Frog development. Reducing testing time in improving the efficiency of many teams working on separate projects. I could do a cost analysis of the savings in developer cost. Also discussing the shortening the go to market time in an emerging market and how important that is.We could add that the networking needs for the VPN deployment was senificate to be able to recreate

Vote Now

Deploying an OpenStack cloud has come a long way since the early days, but even with how far we’ve come a production ready deployment isn’t trivial. There is a lot of thought and planning that needs to be put in place in order to make cloud deployments successful. In this talk, we will discuss the nature of the OpenStack Juju Charms, how to plan for your environment with Juju, and how to build and grow your cloud.

Vote Now

Telecom service providers are in the center of discussion of how to combine the best attributes of clouds, OpenStack, network function virtualization (NFV), software defined networks (SDNs) and virtualized network functions (VNFs), and deployment of those. Toby Ford (ATT), Alan Kavanagh (Ericsson) and John Zannos (Canonical), will discuss the rapidly evolving requirements and solutions that are under consideration. We will discuss how OpenStack provides a common API service framework that can support Telecom needs in NFV and SDN. Some of these needs are unique to Telecom but most are common to the enterprise. Points of discussion will include:Deploying an OpenStack cloud has come a long way since the early days, but even with how far we’ve come a production ready deployment isn’t trivial. There is a lot of thought and planning that needs to be put in place in order to make cloud deployments successful. In this talk, we will discuss the nature of the OpenStack Juju Charms, how to plan for your environment with Juju, and how to build and grow your cloud.

Vote Now

Storage latency is often the enemy of SQL and other Enterprise workloads. Commodity, networked storage used in OpenStack clouds is often in direct contrast to the low-latency needs of enterprise storage. In order to drive more enterprise workloads to OpenStack clouds, a different class of storage may be needed. By providing storage caching at the hypervisor level, using low-latency, high-speed local storage (flash), additional cloud workloads can attain their performance goals. In this session, Canonical and Fusion-io will show how OpenStack storage performance can be accelerated through the use of several storage caching software solutions and flash, or other low-latency local storage. A method for automated deployment and configuration of the caching software for various storage solutions will be detailed. Comparison to conventional storage and caching performance will be outlined.

Vote Now

Enterprise workloads in OpenStack often need high-performance networking. Specific workloads can benefit from accelerated network speeds and lower latency between VMs and their block storage. In this session, Canonical and Mellanox will show how OpenStack network performance can be accelerated through the use of Remote Direct Memory Access (RDMA) over iSCSI (iSER) in 10Gb and 40Gb Ethernet environments. A method for utilizing iSER as the transport layer for VM storage will be described. We also discuss the benefits of iSER in virtualized storage systems and the advantages of using a converged Ethernet infrastructure for large-scale deployments. The session will outline the additional features and performance iSER brings to the deployments using Ubuntu and Juju as their OpenStack deployment tools. The audience will gain valuable insight into the details of configuring iSER and CPU traffic offloads for their storage VMs.

Vote Now

In this presentation you will find answers to the following questions and to other common questions related to the deployment, configuration and use of MongoDB and MySQL databases in OpenStack:

• What is the impact of virtualisation and the OpenStack infrastructure on database performance?
• How can we make MongoDB and MySQL highly available with no data loss and no single points of failure?
• Which storage provides the best tradeoff between availability and performance?
• What is the impact of typical network configurations on database access?
• What do I gain and lose when I use Trove?

We will review and analyse the results of popular benchmarks for MongoDB and MySQL in typical configurations, and some typical architectures implemented by users to achieve optimum performance and high availability.

Vote Now

Based on speaker’s previous experience with large-scale MongoDB deployments the the attemp will be made for OpenStack users to drift from MongoDB to Mysql/Percona and/or PostgreSQL.

It should be no more than 20-30 minutes talk, including questions. Basic pitfalls with MongoDB will be shown, and performance comparisons to other Ceilometer-supported database management systems will be made.

Vote Now

Amazon Aurora has been a major announcement at re:Invent in Las Vegas last November and introduced probably the most innovative relational DBaaS in the market to date.
In OpenStack, we can provide a service that is similar or even better than Amazon Aurora in terms of performance, availability and ease of use.

In this presentation, after e quick review of the features available in RDS Aurora, we will discuss:

• How to implement self-provisioning for your workloads with MySQL
• How to achieve high performance for read and write operations
• How to provide automatic failover in near-realtime with no data loss
• How to implement disaster recovery and multi-region replication
• How to integrate existing monitoring tools and monitoring plugins to your DBaaS architecture

An overview of the self-management tools available for your database service

Vote Now

Public cloud packaging is fairly standardized across the big vendors as Public cloud services are commodity with price being a primary means of assessment. You have to be agile (and brave!) to enter this market. OpenStack and Software Defined Networking (SDN) provide a powerful combination to quickly create unique services at a low cost so you can differentiate your cloud offerings.

Wingu recently launched a public OpenStack cloud targeting large enterprises and managed service providers in South Africa. The business delivers a number of Infrastructure-as-a-Service, Platform-as-a-Service and Software-as-a-Service technologies. Thomas Lee, chief architect of Wingu, will be joined by Sanjay Mishra of Talligent, Canonical, and Juniper, on a panel discussion to go over key aspects of launching a new cloud that can compete both on price and differentiated services.

• The business challenge – Identifying customer pain and differentiating cloud services;
• Architecting a flexible and robust cloud infrastructure using Ubuntu OpenStack;
• Leveraging the benefits of OpenContrail software defined networking;
• Developing rate plans and management reports using OpenBook by Talligent.

Vote Now