Notice: The views and opinions expressed here are collective derived from the members of this ONUG working group and are not the express opinion of any individual or company.
October 8, 2015 Meeting Notes
– Nick & Maxime work on the use cases.
– Stefan document the current state of the industry.
– Brian document the framework
– All work on the benefits.
DRAFT USE CASES
– I want to deploy Salesforce and make it highly available. Instance running in data center A and it is in a standby in data center B. Data center goes down, the orchestrator will reconfigure the environment in seconds and not hours.
– SD-WAN controller and Data center controller. Application that span data center and wide-area.
– Having different controllers, how the different controllers communicate to setup the application dependency map. The dependency map is firewall, wan optimization, switches… Data that we got from different ONUGs, it is 200 days to get a server install in a rack and 42 days to have all the dependency map vs. 15 minutes to setup a VM and X minutes to update the dependency maps.
– This will be to reconfigure all network equipment of an enterprise (switches, routers & access points) because a vendor changed the TCP ports of their application (due to a new patch or a new release) and now you have to update all the QoS of the enterprise.
– Another case could be to configure the IT infrastructure of an enterprise (data centers and offices) to provision a new application or modify an application that has different types of control planes distributed across the enterprise.
– Currently there is no holistic application and service view.
– What are the needs and wants from the community?
September 17, 2015 Meeting Notes
|Brian Hedstrom (Datavision)||Matthew Clark (Hoffman-La Roche)||Stefan Dietrich (Glue Networks)||Charlotte Jolicoeur (ONUG)|
|Jeff Gray (Glue Networks)||David White (Institute of Tech Tallaght Dublin)||Maxime Bugat (Deloitte)||Nick Lippis (ONUG)|
|Rashad Mahmood Khan (Apecon)||Jeannette Tibbetts (ONUG)|
We need to write a white paper. It will 3 or 4 pages very high level. We will take the network monitoring white paper as an example.
The use case will be “Automated workload provisioning”.
Example: An application that spans between 2 data centers. We would envision that the orchestrator will configure the different equipment to provision virtual servers, databases,…, network equipment, bandwidth between data centers.
Maxime will create a “white paper” page and start to draft it.
Brian sent us a (see below). He is going to update it based on today’s call.
We can find open interfaces at https://www.opendaylight.org/lithium.
August 20, 2015 Meeting Notes
Calls will be every two weeks, with the conference being a face-to-face meeting. Collaboration to take place on the Wiki page in between meetings. Promote the WG and invite others to join. The main goal of this group is to write a use case for presentation at the fall ONUG conference. Afterwards, it will begin work on feature verification testing, interoperability and proof of concepts. Now, we will start to define what we need and want.
SDN Federation/Operability Orchestration Use Cases
- Problem being solved
- Lack of interoperability North/South between controllers
- Controllers’ islands
- More and more controllers being developed
- Problem being deferred
- Lack of interoperability East/West between controllers
- Open Networking components
- Any controllers (Storage, WAN, Server, WiFi, LAN…)
- Controller or orchestrator for legacy networking components
- API configuration of controllers
- API calls and handshake between controllers
- Orchestration of control planes of legacy networking elements
- Automatic configuration
- Simpler and faster provisioning
- Capital software instead of capital hardware