Jump to: navigation, search

Difference between revisions of "Tacker/Incubation"

Line 106: Line 106:
 
TODO(to be written)
 
TODO(to be written)
  
=== Team ===
+
= Team =
 
The team has worked for more than 3 months as Neutron subteam.
 
The team has worked for more than 3 months as Neutron subteam.
 
It as independent project out of Neutron started to work since
 
It as independent project out of Neutron started to work since
Line 112: Line 112:
 
From three(XXX) differenct companies(Cisco, Midokura, Intel. XXX)
 
From three(XXX) differenct companies(Cisco, Midokura, Intel. XXX)
  
== Project developers qualifications ==
+
== Team Members ==
 
=== Isaku Yamahata(yamahata on irc) ===
 
=== Isaku Yamahata(yamahata on irc) ===
 
Intel Senior Software Engineer - Core Developer
 
Intel Senior Software Engineer - Core Developer
Line 128: Line 128:
 
xxx(put BIO)
 
xxx(put BIO)
  
=== Hareesh Puthalath ===
+
=== Hareesh Puthalath (hareeshp on irc) ===
 
Cisco - Developer
 
Cisco - Developer
xxx(put BIO)
+
 
  
 
=== Sridar Kandaswamy  ===
 
=== Sridar Kandaswamy  ===

Revision as of 09:36, 25 November 2014

This page is Draft under review

Project codename

Tacker

Trademarks

(Does this project name, codename or contents use any trademarks, and if so, who owns them? List the codenames or other marks for which a trademark search should occur.)

No known trademarks.


Summary

Tacker is a servicevm/device life cycle management for openstack clouds


Parent Program name and PTL

Neutron networking/Kyle Mestery


Mission statement

provide common framework/ground work for life cycle management of ServiceVMs and virtual/physical services which conforms ETSI NFV.


Mission

establish/encourage ecosystem of virtual/physical appliance on OpenStack. To make it easy for appliance providers to integrate their appliances (services implemented in virtual machines (a.k.a. ServiceVM) or physical hardware) with OpenStack by providing unified interface for life cycle management of VMs/Services which is capable of multiple service/vendor.

Detailed Description

(Background: this project is spin-out project from Neutron)

It is a quite common requirement to run services in virtual machines aka ServiceVM. So far each OpenStack service(Especially Neutron case) implemented their own life cycle management of ServiceVMs/services and hardware. Preparing/pooling/scheduling VMs/services and so on. It resulted in duplicated work and rises the bar for appliance provide to integrate appliances with OpenStack.

This project introduces a new service for managing servicevm/device. Its responsibility is to:

  1. manage VMs/devices/services
  2. control the allocation of processing capacity in such devices As such the servicevm/device manager mainly serves other projects. Another natural responsibility of this service would be to keep track of and store the physical topology that the devices are part of.
  3. pluggable for each VMs/physical devices/services


As ETSI NFV is an important use case, NFV conformance is in the scope of this project. This project services as some components of NFV MANO architecture. http://www.ietf.org/proceedings/89/slides/slides-89-opsawg-7.pdf

  • corresponding component:
    • VNFM(VNF Manager)
  • corresponding interfaces:
    • VNF lifecycle management interface
    • VNF lifecycle changes notifications interface
    • VNF configuration interface


The relationship of NFV team and this project:

This project serves as (sub)component(s) necessary for NFV in openstack. The team will cooperate with NFV team.

Basic roadmap for the project

from Near term to middle term (Note: Addressing missing features for Neutron/Nova: some features are missing in Neutron/Nova. Those features can be addressed independently of this project.) Design/patch review process will follow Neutron style as this project spun out of Neutron.

  • terminology conversion. including NFV conformance/nomenclature
  • produce a minimul working code as an independent service(out of Neutron. So far there were two implementations within Neutron). merge floating around multiple implementations.
  • design review process: gerrit, wiki, google-doc
  • API/data model design
  • Actual coding(server, python client)
  • establish testing
  • Corporate with Neutron driver for Network service in VM (e.g. router vm, firewall vm) that interacts with Tacker
  • make use of currently missing Neutron feature like service insertion

At first temporal shim code will be needed in Neutron for service insertion, traffic steering and so on. At this moment those features are under development. Once such features are stabilized and public RestAPI is provided, convert the implmentation to use RestAPI and remove the temporal shim code.

  • consider generic VM/service scheduling

Location of project source code

TODO:xxx

Programming language, required technology dependencies

language: Python Dependencies: alembic, eventlet, sqlalchemy, pbr, xxx (??? pecan, jsonschema, ...)

Is project currently open sourced? What license?

yes and apache 2.0


Level of maturity of software and team

Software

TODO(to be written)

Team

The team has worked for more than 3 months as Neutron subteam. It as independent project out of Neutron started to work since May 2014. From three(XXX) differenct companies(Cisco, Midokura, Intel. XXX)

Team Members

Isaku Yamahata(yamahata on irc)

Intel Senior Software Engineer - Core Developer

Isaku is a developer who has been involved with OpenStack since Diablo cycle. He has contributed the first version of boot-from-volume to Nova, Ryu plugin to Neutron. Recently he's mainly been contributing to Neutron and servicevm project.

Bob Melander(bmelande on irc)

Cicco - Core Developer xxx(put BIO)

Stephen Wong(s3wong on irc)

Midokura - Core Developer xxx(put BIO)

Hareesh Puthalath (hareeshp on irc)

Cisco - Developer


Sridar Kandaswamy

Cisco - Developer xxx(put BIO)

Yong sheng gong (gongysh on irc)

Unitedstack - Developer

Karthik Natarajan (natarajk on irc)

Brocade - Developer

Kyle Mestery (mestery in IRC)

Cisco - Developer

Balaji Padnala (balajip on irc)

Freescale

Sridhar Ramaswamy (SridharRamaswamy on irc)

Brocade - Developer

Infrastructure requirements (testing, etc)

no additional requirements to the current OpenStack Infrastructure

Have all current contributors agreed to the OpenStack CLA?

yes