OpenStack Summit | Austin, TX
You can build your OpenStack
and consume it too
Copyright © PLUMgrid, Inc. 2011-2016
Introduction
Speaker(s)
2
Sr Director Product & Solution Marketing, PLUMgrid
Valentina Alaria
Part of the OpenStack community for 5+ years and has worked with 100s of users throughout their journey of
learning, designing and deploying OpenStack-based cloud solutions. A product innovation strategist and
technology evangelist, Valentina has 10+ years experience with Cloud and Datacenter Infrastructure and has
been involved with SDN since the early days throughout her endeavors at PLUMgrid, Nicira and Cisco.
Sr Director Product Management, Rackspace
Bryan Thompson
Leads the product team for Rackspace Private Cloud. Prior to joining Rackspace, Bryan served as VP of
Product management at Tier 3 and held product and technology leadership roles at Limelight Networks and
Amazon.com. He brings over fifteen years of experience in technology building, deploying and operating
solutions for the enterprise.
To adopt or not to adopt
3
Copyright © PLUMgrid, Inc. 2011-2016
Complexity: The Primary Barrier to OpenStack Adoption
4
• OpenStack is not a “Product”
• Massive number of potential
configurations
• Nova alone has 8 different
virtualization drivers and over 700
unique configurable attributes
• 48 different Neutron plugins
• 69 different Cinder drivers
Copyright © PLUMgrid, Inc. 2011-2016
Unmanaged Managed
The Cloud Market is Evolving
5
Cloud
Market
Copyright © PLUMgrid, Inc. 2011-2016
Unmanaged: Make your own car
6
Copyright © PLUMgrid, Inc. 2011-2016
Managed: Get the complete car
7
Copyright © PLUMgrid, Inc. 2011-2016
Road to Cloud Networking
#1
DEPLOYMENT
ROADBLOCK
75
Open
Tickets
4
Months
Delay
#1
DEPLOYMENT
ENABLER
0
Open
Tickets
0
Weeks
Delay
Copyright © PLUMgrid, Inc. 2011-2016
• Started with the Folsom release
• Provide Network as a Service
• Provide Operator & Tenants ability to create and
offer rich network topologies and configure
advanced policies
• Offer a technology agnostic layer while enabling
vendor extensions
• Support for advanced services
Why Neutron?
Compute Storage
Network
9
Copyright © PLUMgrid, Inc. 2011-2016
LBaaS
Agent
FWaaS
Agent
DHCP
Agent
L3 Agent
AgentAgent
Agent
Agent
Agent
Agent
Architecture Challenges: Neutron & Network Nodes
Neutron
ML2/OVS plugin
Network
Nodes
Nova Glance Swift CinderServices Neutron
Framework
Placement of these components
is critical; They are in data path
and become bottlenecks
Advanced Services run on
dedicated nodes. Limited HA.
Creation of new tenants requires careful
sizing of components to maintain
performance level
VM traffic flow can be handled in kernel, in
local user space or in network nodes with
different performance level
10
Rackspace + PLUMgrid Joint Solution
11
Copyright © PLUMgrid, Inc. 2011-2016
Why Rackspace and PLUMgrid for OpenStack
12
We make
OpenStack
SIMPLE
We eliminate complexities
by delivering OpenStack
“as a service” and backing
it with an industry-leading
99.99% uptime SLA.
Copyright © PLUMgrid, Inc. 2011-2016
Why Rackspace and PLUMgrid for OpenStack
Business Value
13
1
2
3
4
Single point of contact for all your OpenStack operations
Tight integration, validation and certification
Combined Expertise: Design, Deploy, Operate
One Support Number + Commitment to Excellence
Copyright © PLUMgrid, Inc. 2011-2016
Why Rackspace and PLUMgrid for OpenStack
Technical Value
14
1
2
3
4
Built-in Automation
Enterprise-grade Architecture
Secured & Auditable Platform
Faster time to production
Behind the curtains
15
Copyright © PLUMgrid, Inc. 2011-2016
Rackspace Private Cloud
16
Superior model:
OpenStack-as-a-
Service
(24x7 monitoring,
maintenance, upgrades)
World’s most
reliable
OpenStack:
99.99% uptime
SLA
OpenStack
everywhere: deploy in a
Rackspace
data center, customer data
center or 3rd party location
The most OpenStack operational
experience in the world, by a factor of
100x: we’ve successfully scaled OpenStack to
1,000’s of nodes and operate OpenStack clouds for
some of the largest companies in the world
A single platform for
managing containerized
applications, virtualized
applications and bare metal
applications across private, public
and hybrid clouds
Copyright © PLUMgrid, Inc. 2011-2016
RPC VERSION 12 SOFTWARE – Liberty RELEASE
RPC V12 includes the following OpenStack services:
• Compute (Nova)
• Object Storage (Swift)
• Block Storage (Cinder+Ceph)
• Networking (Neutron+ONS)
• Dashboard (Horizon)
• Image Service (Glance)
• Identify Service (Keystone)
• Orchestration (Heat)
RPC V12 Advantages Include:
• Automation – Deploy production-ready
application stacks in minutes with RPC
solution templates and use our DevOps
services to unlock the power of your cloud by
treating your infrastructure as code.
• Enterprise-Grade Architecture – RPC
includes an industry-leading 99.99%
OpenStack API uptime guarantee, is designed
to scale to hundreds of nodes, delivers HA via
a four-node control plane and uses Linux
containers for service deployment
• Security & Compliance – RPC provides a
secure, dedicated environment that includes
two physical firewalls and audit logs for
compliance reporting
Copyright © PLUMgrid, Inc. 2011-2016
PLUMgrid – Comprehensive Networking Offering
Extensive software-only SDN and NFV solution for OpenStack® Clouds
18
Top Security & Compliance Support with
built-in isolation, micro-segmentation via
Virtual Domains & BYO service
Unmatched Operational tools with
proactive visibility & analytics (powered by
CloudApex)
Virtual networks provisioned by users
Scalable, distributed & highly available
architecture enables Production
Deployments
Copyright © PLUMgrid, Inc. 2011-2016
PLUMgrid ONS Components
Internet
IO Visor Gateway
IO Visor Edges (Compute Nodes)
PLUMgrid Directors
VXLAN-based
Overlay
19
PLUMgrid CloudApex & OpsVM
Your Journey to Private Cloud
20
Copyright © PLUMgrid, Inc. 2011-2016
A Comprehensive Offering
Enablement Services, Training, Operations and Support
Support:
Operating Your Cloud
Cloud Enablement:
Optimizing Your Cloud
Training: Take Full
Advantage of Your Cloud
• RPC Managed Services
• Dedicated OpenStack Engineers
• DevOps Automation Services
• Engagement Guide
• Application Modernization
• Agile Methodology
• Authentication & Federation
• Big Data
• Cloud Optimization
• RPC Cloud User Training
• OpenStack Fundamentals
• PLUMgrid ONS Training
Cloud Design:
Planning Your Cloud
• Workload Definition
• Application Profiling
• Sizing and Scale Planning
• Deployment
Copyright © PLUMgrid, Inc. 2011-2016
Deployment Reference Architecture
22
Compute Node Compute Node
Fabric
eth1 eth1 eth1eth0 eth0eth1
Southbound
VLAN(s)
eth0
Controller #2
eth0
Storage
Controller #3
eth1
eth1DEPLOY
NODE
eth0eth0
LXC
VM BAREMETAL
eth0
eth0
eth1
eth1
BAREMETAL OR LXC
LCM + Deployment Node
Controller #1
Storage Node
eth1 eth0
Management
External
Internet
Intranet
LCM
Director 1
LXC BAREMETAL OR LXC
Director 2
LXC BAREMETAL OR LXC
Director 3
Gateway
#2
Gateway
#1
So: What can I do with this?
23
Copyright © PLUMgrid, Inc. 2011-2016
Deployment Use Cases
Your Private Cloud
• NAT at scale
• Web Application
Template
• High Availability
• Security
• Performance
• Partitioning of
resources with
no physical
network
changes
• Multi-cloud
support
• High Availability
• Low latency
performance
• L2 extensions,
DHCP, DNS
• On-demand
Virtual Domains
• Arbitrary
Topologies
• No changes in
Physical
Network
• Security,
isolation &
partitioning
• Any software
Any where
• CRM, ERP
• ITaaS
• Security
• Performance
• Multi-cloud
Support
PaaS SaaS E-Commerce Hadoop
As a Service
Communications
as a Service
THANK YOU!

You Can Build Your OpenStack and Consume it Too

  • 1.
    OpenStack Summit |Austin, TX You can build your OpenStack and consume it too
  • 2.
    Copyright © PLUMgrid,Inc. 2011-2016 Introduction Speaker(s) 2 Sr Director Product & Solution Marketing, PLUMgrid Valentina Alaria Part of the OpenStack community for 5+ years and has worked with 100s of users throughout their journey of learning, designing and deploying OpenStack-based cloud solutions. A product innovation strategist and technology evangelist, Valentina has 10+ years experience with Cloud and Datacenter Infrastructure and has been involved with SDN since the early days throughout her endeavors at PLUMgrid, Nicira and Cisco. Sr Director Product Management, Rackspace Bryan Thompson Leads the product team for Rackspace Private Cloud. Prior to joining Rackspace, Bryan served as VP of Product management at Tier 3 and held product and technology leadership roles at Limelight Networks and Amazon.com. He brings over fifteen years of experience in technology building, deploying and operating solutions for the enterprise.
  • 3.
    To adopt ornot to adopt 3
  • 4.
    Copyright © PLUMgrid,Inc. 2011-2016 Complexity: The Primary Barrier to OpenStack Adoption 4 • OpenStack is not a “Product” • Massive number of potential configurations • Nova alone has 8 different virtualization drivers and over 700 unique configurable attributes • 48 different Neutron plugins • 69 different Cinder drivers
  • 5.
    Copyright © PLUMgrid,Inc. 2011-2016 Unmanaged Managed The Cloud Market is Evolving 5 Cloud Market
  • 6.
    Copyright © PLUMgrid,Inc. 2011-2016 Unmanaged: Make your own car 6
  • 7.
    Copyright © PLUMgrid,Inc. 2011-2016 Managed: Get the complete car 7
  • 8.
    Copyright © PLUMgrid,Inc. 2011-2016 Road to Cloud Networking #1 DEPLOYMENT ROADBLOCK 75 Open Tickets 4 Months Delay #1 DEPLOYMENT ENABLER 0 Open Tickets 0 Weeks Delay
  • 9.
    Copyright © PLUMgrid,Inc. 2011-2016 • Started with the Folsom release • Provide Network as a Service • Provide Operator & Tenants ability to create and offer rich network topologies and configure advanced policies • Offer a technology agnostic layer while enabling vendor extensions • Support for advanced services Why Neutron? Compute Storage Network 9
  • 10.
    Copyright © PLUMgrid,Inc. 2011-2016 LBaaS Agent FWaaS Agent DHCP Agent L3 Agent AgentAgent Agent Agent Agent Agent Architecture Challenges: Neutron & Network Nodes Neutron ML2/OVS plugin Network Nodes Nova Glance Swift CinderServices Neutron Framework Placement of these components is critical; They are in data path and become bottlenecks Advanced Services run on dedicated nodes. Limited HA. Creation of new tenants requires careful sizing of components to maintain performance level VM traffic flow can be handled in kernel, in local user space or in network nodes with different performance level 10
  • 11.
    Rackspace + PLUMgridJoint Solution 11
  • 12.
    Copyright © PLUMgrid,Inc. 2011-2016 Why Rackspace and PLUMgrid for OpenStack 12 We make OpenStack SIMPLE We eliminate complexities by delivering OpenStack “as a service” and backing it with an industry-leading 99.99% uptime SLA.
  • 13.
    Copyright © PLUMgrid,Inc. 2011-2016 Why Rackspace and PLUMgrid for OpenStack Business Value 13 1 2 3 4 Single point of contact for all your OpenStack operations Tight integration, validation and certification Combined Expertise: Design, Deploy, Operate One Support Number + Commitment to Excellence
  • 14.
    Copyright © PLUMgrid,Inc. 2011-2016 Why Rackspace and PLUMgrid for OpenStack Technical Value 14 1 2 3 4 Built-in Automation Enterprise-grade Architecture Secured & Auditable Platform Faster time to production
  • 15.
  • 16.
    Copyright © PLUMgrid,Inc. 2011-2016 Rackspace Private Cloud 16 Superior model: OpenStack-as-a- Service (24x7 monitoring, maintenance, upgrades) World’s most reliable OpenStack: 99.99% uptime SLA OpenStack everywhere: deploy in a Rackspace data center, customer data center or 3rd party location The most OpenStack operational experience in the world, by a factor of 100x: we’ve successfully scaled OpenStack to 1,000’s of nodes and operate OpenStack clouds for some of the largest companies in the world A single platform for managing containerized applications, virtualized applications and bare metal applications across private, public and hybrid clouds
  • 17.
    Copyright © PLUMgrid,Inc. 2011-2016 RPC VERSION 12 SOFTWARE – Liberty RELEASE RPC V12 includes the following OpenStack services: • Compute (Nova) • Object Storage (Swift) • Block Storage (Cinder+Ceph) • Networking (Neutron+ONS) • Dashboard (Horizon) • Image Service (Glance) • Identify Service (Keystone) • Orchestration (Heat) RPC V12 Advantages Include: • Automation – Deploy production-ready application stacks in minutes with RPC solution templates and use our DevOps services to unlock the power of your cloud by treating your infrastructure as code. • Enterprise-Grade Architecture – RPC includes an industry-leading 99.99% OpenStack API uptime guarantee, is designed to scale to hundreds of nodes, delivers HA via a four-node control plane and uses Linux containers for service deployment • Security & Compliance – RPC provides a secure, dedicated environment that includes two physical firewalls and audit logs for compliance reporting
  • 18.
    Copyright © PLUMgrid,Inc. 2011-2016 PLUMgrid – Comprehensive Networking Offering Extensive software-only SDN and NFV solution for OpenStack® Clouds 18 Top Security & Compliance Support with built-in isolation, micro-segmentation via Virtual Domains & BYO service Unmatched Operational tools with proactive visibility & analytics (powered by CloudApex) Virtual networks provisioned by users Scalable, distributed & highly available architecture enables Production Deployments
  • 19.
    Copyright © PLUMgrid,Inc. 2011-2016 PLUMgrid ONS Components Internet IO Visor Gateway IO Visor Edges (Compute Nodes) PLUMgrid Directors VXLAN-based Overlay 19 PLUMgrid CloudApex & OpsVM
  • 20.
    Your Journey toPrivate Cloud 20
  • 21.
    Copyright © PLUMgrid,Inc. 2011-2016 A Comprehensive Offering Enablement Services, Training, Operations and Support Support: Operating Your Cloud Cloud Enablement: Optimizing Your Cloud Training: Take Full Advantage of Your Cloud • RPC Managed Services • Dedicated OpenStack Engineers • DevOps Automation Services • Engagement Guide • Application Modernization • Agile Methodology • Authentication & Federation • Big Data • Cloud Optimization • RPC Cloud User Training • OpenStack Fundamentals • PLUMgrid ONS Training Cloud Design: Planning Your Cloud • Workload Definition • Application Profiling • Sizing and Scale Planning • Deployment
  • 22.
    Copyright © PLUMgrid,Inc. 2011-2016 Deployment Reference Architecture 22 Compute Node Compute Node Fabric eth1 eth1 eth1eth0 eth0eth1 Southbound VLAN(s) eth0 Controller #2 eth0 Storage Controller #3 eth1 eth1DEPLOY NODE eth0eth0 LXC VM BAREMETAL eth0 eth0 eth1 eth1 BAREMETAL OR LXC LCM + Deployment Node Controller #1 Storage Node eth1 eth0 Management External Internet Intranet LCM Director 1 LXC BAREMETAL OR LXC Director 2 LXC BAREMETAL OR LXC Director 3 Gateway #2 Gateway #1
  • 23.
    So: What canI do with this? 23
  • 24.
    Copyright © PLUMgrid,Inc. 2011-2016 Deployment Use Cases Your Private Cloud • NAT at scale • Web Application Template • High Availability • Security • Performance • Partitioning of resources with no physical network changes • Multi-cloud support • High Availability • Low latency performance • L2 extensions, DHCP, DNS • On-demand Virtual Domains • Arbitrary Topologies • No changes in Physical Network • Security, isolation & partitioning • Any software Any where • CRM, ERP • ITaaS • Security • Performance • Multi-cloud Support PaaS SaaS E-Commerce Hadoop As a Service Communications as a Service
  • 25.

Editor's Notes

  • #5 Complexity and perceived immaturity are the two primary barriers to OpenStack adoption. Complexity: The infinite number of configurations and subsequent lack of operational experience and OpenStack expertise hinder adoption. Perceived Immaturity: While it has matured over the last 5 years and 12 releases, there is still some residual perception that OpenStack isn’t reliable or ready for production workloads.
  • #10 Openstack networking started with nova-network: a very basic way of using VLANs to interconnect VMs. Very limited.
  • #14 #1: from initial contract through monthly billing and updates
  • #15 HA, hardening, scale
  • #18 Insert information about the foundation we laid with Version 9 --- designed for production-grade in the enterprise, with scalability, availability and automation in mind. Moved to a container-based architecture. Using Linux containers (LXC) to deploy the OpenStack services which isolates the services and improves cross-service interconnectivity. This allows customers to independently scale each service and enables us to provide seamless, in-place upgrades from one release to another. Moved to a 3-node MariaDB Galera cluster with quorum. This allows for better handling of single-node failures and recovery of a lost node. In the previous architecture we encountered a number of split-brained scenarios with both nodes thinking they were master. Moved RabbitMQ to a 3-node cluster with quorum. Provides the same improvement to fault tolerance. Allows us to utilize OpenStack’s ability to connect to multiple Rabbit nodes along with HA queues – again, providing a more fault tolerant and scalable environment as these clusters can be easily scaled where as we were stuck with the pair in previous architectures. Replaced Chef with Ansible for automated cloud deployment and configuration management. With the move to containers, the Chef Package/Service model was much heavier than we needed. Ansible provides us an automation framework great at orchestrating tasks across multiple hosts (sequentially or in parallel). We use Ansible to lay down containers and manage the installation of source code. Added robust audit logs for compliance reporting – added a controller node for Logstash, ElasticSearch and Kibana – this is our logging server which centralizes logging services to help provide customers and our support team with event services data Added Heat orchestration – enables customers to automatically provision infrastructure, services and applications Insert additional enhancements from Version 10