A session in the DevNet Zone at Cisco Live, Berlin. There are several new OpenStack projects/services that are built on core OpenStack infrastructure services. This session will first briefly discuss the changes introduced for the project governance structure in OpenStack. Subsequently, the focus of the presentation will be to provide feature and architecture details on few of the new projects and services in OpenStack. These will include Trove-Database Service, Sahara-Dataprocessing Service, Congress - Policy Service and Magnum -- Container Service. A summary of other OpenStack related services will also be provided.
6. OpenStack Progress
Austin – Oct 2010
Bexar – Feb 2011
Cactus– April 2011
Diablo – Sept 2011
Essex – April 2012
Folsom– Sept 2012
Grizzly– April 2013
Havana – Oct 2013
IceHouse– April 2014
Juno – Oct 2014
Kilo – May 2015
130 contributors
30 new features
2010
2011
2012
2013
2014
Started with Compute
and Storage service
12th OpenStack release
1933 contributors
760 new features
8300 bugs fixed
164 companies
Liberty – Oct 2015
24,000 people
495 companies
Mitaka – April 2016
2015
Newton – Oct 2016
8. • Database as a Service - Automating
complex database administrative
tasks – Deployment, Configuration,
Scaling, HA
• Single Admin Tenant Database per
Nova instance
• Datastore type - Relational and Non-
Relational
• Pluggable – Support for MySQL,
PostgreSQL, NoSQL – Cassandra,
MongoDB, Couchbase, Redis
• Integration with other projects –
Designate, Heat, Neutron
Trove
API Functionality
Management Create/Delete/Show/List Database
instance, Database, Users, Flavors
Security Support for Security Groups, No SSH by
default
Configuration
groups
Support for user defined configuration
settings (MySQL, MongoDB) – Eg:
max_connections, Buffer pool size etc
Backups Support for full backups and incremental
backups using Swift (MySQL,
Cassandara, Couchbase)
Replication Async mysql master-slave replication from
snapshot of master
Clustering Support for shards - three member replica-
sets (MongoDB)
10. • Cluster provisioning - Create and Manage Hadoop clusters
• Node Group Templates – Defines instances/nodes within a
cluster that will each run selected hadoop processes and store
data
• Plugins – Responsible for provisioning Hadoop cluster – Vanilla,
Hortonworks, MapR, Clouder, Spark
• Cluster Templates – What Node Groups to be included and
how many instances to be created in each
• Anti Affinity Groups – Processes may not be launched more than
once on a single host
• Cluster – Represents a Hadoop cluster run using Cluster
Template
• Image Registry – Used to provide additional information about
images using tags
• Cluster Scaling - Change instances in a existing Node group or add
new Node groups
Sahara
• Configure HDFS and MapReduce
parameters at Node and Cluster level
• Integration with Cinder, Swift, Neutron, Heat
API Functionality
Data
Sources
Stores URL which designates
the location of input and
output data
Job
Binaries
URL to script or Jar file stored
in internal DB or Swift
Jobs Specifies the job and lists all
individual Job Binary objects
required for execution
Job
Execution
Monitor and Manage a Job
executed/launched on a
Cluster
12. Ironic
• Service for Bare metal management
• Ironic API – RESTful API service
• Ironic Conductor – Interacts with hardware;
asynchronous handling of both requested and periodic
actions
• Ironic Python Agent – utility service temporarily booted
on machines to provide remote access to hardware for
provisioning and management
• Ironic Drivers – Communicate with hardware devices
• Nova driver
• Using Nova API can provision bare metal servers
• Based on Open technologies
• DHCP, TFTP, PXE - In a PXE environment, TFTP is
used to download NBP over the network using
information from the DHCP server.
• Ironic API – Chassis, Drivers, Links, Nodes, Ports
14. • Nova Container virtualization drivers - Docker, LXC, OpenVZ,
ZeroVM
• Heat resource for Docker
• Container as a Service (“Nova of containers”) - Provide a REST
API for Container management
• Provide app isolation, portability, manageability with containers
• Containers in VMs, Bare Metal, Containers
• Resources
• Container – a container
• Node – a bare metal or virtual machine where work executes
• Bay – Collection of nodes where work is scheduled
• Service – a port to Pod mapping
• Pod – a collection of containers running on one physical or
virtual machine
Magnum
Kubernetes
Endpoint
Docker
Endpoint
Mesos
Endpoint
Magnum
Launch instances with
Agent for hosting
Containers
Operation on Service
and Pod objects
Operation on Container
object
• Integration with Kubernetes, Docker,
Mesos
• Companies – Rackspace, RedHat,
Cisco and others
16. • Containerization of OpenStack services - All "core" OpenStack
services implemented as micro-services in Docker containers.
• Technology
• Dockers – Deploying containers and managing images
• Ansible – Orchestration tool for multi node deployment
• Jinja2 – Templating language for Python
• Developing and Deploying OpenStack service using Kolla
• Supports deployment from binary and source across multiple
distributions - CentOS, Debian, Fedora, Oracle Linux, RHEL,
and Ubuntu
• AIO and multinode deployment using Ansible with n-way active
high availability.
• Development environments using Heat, Vagrant, or bare-metal.
Kolla
Node
Keystone
Nova
Glance
Neutron
Horizon
Horizon
….
17. • OpenStack services ecosystem is
expanding
• Lots of opportunities to contribute and
influence the community
Summary