Grounded in the Cloud
cancel
Showing results for 
Search instead for 
Did you mean: 

OpenStack Summit Sessions: User Stories

Stephen_Spector

As mentioned in a previous post, the OpenStack community is now voting on submitted talks for the upcoming Summit in Paris.  I have broken the submissions down into separate blogs to make it easier to find the available HP Helion submitted talks for your voting consideration.

 

USER STORIES

Deploying and running HP’s Platform Services as in-cloud workloads on HP Helion Clouds

In this session we present architecture designs and trade-offs that were considered when we decided to run the Platform Services (DNSaaS, DBaaS, etc.) as in-cloud workloads as part of the HP Helion OpenStack Editions.

Our OpenStack architecture is from a deployment perceptive divided up in four separate layers: seed, under-cloud, over-cloud and in-cloud. The in-cloud layer has for security reasons a high level of isolation from the over-cloud and under-cloud given that tenants running on VMs in-cloud are not in most cases trusted tenants.

 

We describe the operator scale and QoS requirements which guided our strategy around the deployment and management of the services and how we designed a single framework that fit the existing  as well as future Platform Services.

 

This single framework ensures that the in-cloud services follow best practices with regards to resiliency and HA, deployment using Heat, and integration with monitoring, metering, a centralized logging. We also present how we use a Neutron Provider Networks to allow each of the services access to the over-cloud services. 

 

NFS based Nova Instance booth process optimization  

In a given public or private cloud, the optimization of the instance boot process is one of the critical challenge that cloud service providers are facing today. The same problem exists for the OpenStack cloud service providers as well. Usually the instance are booted with either directly from the given image or from the bootable volume, which is already loaded with given image to be booted. In earlier case, the image is copied from glance image store to hypervisor image cache via glance public REST API and nova-compute processes and then finally hypervisor boot the instance from the image available in the local image cache. And in the latter case, image is copied in to the volume from the glance image store via glance public REST API and cinder-volume processes. In both of these cases, image is copied over multiple processes and introduces the time delay in overall boot process. 

 

This problem is investigated and found a solution by using the NFS as the glance image store and mount that NFS image store in all hypervisors used in the cloud. Here nova-compute and cinder-volume processes will trigger the image copying processes to copy the image directly from the glance image store to the hypervisor and those processes does not involve in image copy operation.  In POC it is revealed that the boot process got speedup by 62%.

 

Taking advantage of our OpenStack cloud for a well-established R&D

In our companies, we are re-inventing the way we develop, build, test, deploy and operate our products, as well as our IT assets. We are figuring out how to best take advantage of our private and public OpenStack powered clouds. Many of our R&D departments have been established for some times, and they also are looking at the new private and public cloud as a way to increase their velocity. But are we, in IT, only supposed to provide tons of VMs or containers without much more added value than that? 

 

In this session, Olivier from HP R&D IT and Jérôme from Alcatel-Lucent R&D IT will share where they are in their journey in powering each companies R&D departments with an OpenStack cloud which provides not only infrastructure as a service, but also development and test as a service, platform as a service - all of this in a very practical and hands-on way, from practitioners.

 

In our companies, we are re-inventing the way we develop, build, test, deploy and operate our products, as well as our IT assets. We are figuring out how to best take advantage of our private and public OpenStack powered clouds. Many of our R&D departments have been established for some times, and they also are looking at the new private and public cloud as a way to increase their velocity. But are we, in IT, only supposed to provide tons of VMs or containers without much more added value than that? 

 

In this session, Olivier Jacques from HP R&D IT and Jérôme Fillière from Alcatel-Lucent R&D IT will share where they are in their journey in powering each companies R&D departments with an OpenStack cloud which provides not only infrastructure as a service, but also development and test as a service, platform as a service - all of this in a very practical and hands-on way, from practitioners.

 

HP Helion’s results after 6 months of deploying OpenStack Swift – the good, the bad, the ugly  

The HP Helion IT team will take you through a detailed overview of their OpenStack Swift object storage deployment, from why they chose Swift to build their own private cloud, to what use case they have deployed on, how easy it was to deploy and what the results have been so far.

 

Cloud Journey Story-Telling: An enterprise success story on OpenStack, from multiple angles  

An enterprise success story on OpenStack, from multiple angles.

 

Today our printing business operates a portfolio of web-services, iterating and evolving at high tempo, serving hundreds of thousands of end customers. Our business services are deployed as 85 discrete stacks comprised of >1,500 VMs, managed with grace by a team of just ten. We will share how we approached our integration with OpenStack, what challenges we confronted along the journey (reliability, consistency, scale-out), how we deliberated build-versus-buy decision-making for tools and service-management, and a tell-all review of the many unforeseen surprises we faced.

 

Throughout our stand-up and scale-out work we committed to these three objectives:

  • Increase business agility: deliver to the greater organization an      on-tap capacity to iterate and course-correct services’ features and      strategies at high tempo;
  • Increase operations efficiency: Manage more stacks with zero team      growth, commit more time to R&D cooperation (guiding the next      applications’ roadmaps) while having to give less cycles to operational      maintenance;
  • Grow our cloud footprint while aggressively managing and optimizing      costs

Our talk will recount the eight months’ project we undertook in 2013-2014, sprinting from a standing-start with no OpenStack experience or operations, to our 1,500-VM steady-state and growing portfolio of customer-facing web-services. We explain how we navigated between requirements of our peer app developers, our operations sysadmins, schedule pressure, and our desire to transform processes to adopt as much end-to-end automation and ‘infrastructure-as-code’ as our partners and the cloud could handle.

 

We will discuss integration with OpenStack public cloud (HP) from our perspective as an external customer, the central role that Heat plays in our orchestration, development and maintenance of “glue code” for certain cloud APIs, of pervasive configuration-management across applications running atop the cloud (we use Puppet), business process monitoring, logs-transparency, telemetry, end-to-end continuous-delivery automation for cloud-resources and business packages, and high-level status reporting and dashboards. We will also talk about keeping an eye on developments and open bugs in OpenStack ‘up-stream’ while not actually being OpenStack developers. We won’t claim to have a universal recipe but in sharing our business’s “cloud journey” we can share practical lessons-learned and recommendations to other enterprise OpenStack adopters operating in the user/app-developer domain.

 

 

 

Senior Manager, Cloud Online Marketing
0 Kudos
About the Author

Stephen_Spector

I manage the HPE Helion social media and website teams promoting the enterprise cloud solutions at HPE for hybrid, public, and private clouds. I was previously at Dell promoting their Cloud solutions and was the open source community manager for OpenStack and Xen.org at Rackspace and Citrix Systems. While at Citrix Systems, I founded the Citrix Developer Network, developed global alliance and licensing programs, and even once added audio to the DOS ICA client with assembler. Follow me at @SpectorID

Events
28-30 November
Madrid, Spain
Discover 2017 Madrid
Join us for Hewlett Packard Enterprise Discover 2017 Madrid, taking place 28-30 November at the Feria de Madrid Convention Center
Read more
See posts for dates
Online
HPE Webinars - 2017
Find out about this year's live broadcasts and on-demand webinars.
Read more
View all