5:30 - 6:00pm: Social networking, pizza, drinks
6:00pm - 7:15pm
Talk & demo: Container networking and Service meshes using Kubernetes, Calico and istio
Speaker: Cody McCain - Solutions Architect, Tigera
7:15pm - 8:00pm
Talk & demo: Short update on container platform solutions using Kubernetes, Contiv, nginx, aci & envoy
Speaker: Sanjeev Rampal, Principal Engineer, Cisco Systems
Online Meeting Info:
To join the Meeting:
To join via phone :
1) Dial:[masked] (United States) [masked] (US Toll Free) (see all numbers - https://www.redhat.com/en/conference-numbers )
2) Enter Conference ID :[masked]
Join us for this session jointly organized with the Triangle Kubernetes and Openshift Meetup (https://www.meetup.com/Triangle-Kubernetes-Meetup/events/240391989/)
Details are here:
5:45-6:15pm: Pizza, drinks and Networking
6:30-7:30: main talk
Speaker: Andre Fredette. SDN Software Architect, Red Hat.
Title: Using OpenDaylight as an SDN Controller for OpenStack Deployments
NetVirt is the central component of a complete open source network virtualization solution for OpenStack. NetVirt works together with Open vSwitch and hardware gateways in the dataplane to control a virtualized multi-tenant network for OpenStack deployments. This solution includes fully distributed layer 2 and layer 3 switching, and provides scalability and resiliency with OpenDaylight controller node clustering. NetVirt is also used as part of the Open Platform for NFV (OPNFV) project to support Service Function Chaining (SFC) in OpenStack deployments.
In this talk, we will provide an overview of OpenDaylight and the OVSDB NetVirt project, and describe how it is used in support of OpenStack and OPNFV.
Speakers: Prayson Pate, CTO, Overture
Billy McFall, Staff Engineer, Overture
Ian Miller, Staff Engineer, Overture
Title: "No Shortcuts Needed on the Drive to Pure-Play Virtualization"
The move to cloud and virtualization requires a move from closed boxes to open solutions, providing a golden opportunity for suppliers of hardware and software components. This change is also causing heartburn for systems suppliers who fear letting go of the closed model. Prayson, Billy and Ian will discuss how a pure-play model of virtualization can enable innovation through systems built from ecosystems of components, and how virtualized systems can meet the performance requirements of services – today and tomorrow.
Part 1: What is pure-play, and how does it enable innovation?
Prayson will define pure-play virtualization, describe its benefits, and discuss a real-world deployment of NVF on the customer site by Masergy.
Part 2: Implementing the vision
Implementing pure-play meant solving some tough problems. Billy and Ian will discuss these topics:
Implementing carrier Ethernet functionality in software, including functions that are sensitive to timing variation e.g. QoS and Y.1731 Service OAM
Achieving high packet throughput on a low-cost Atom (not Xeon) server, including measured results.
Maintaining compatibility with today’s VNFs as well as OpenStack and its Neutron plugin
Architecting for the future including higher speeds and technologies like containers and overlay networks
Pizza, drinks and more!
(1) Cloud: Its all about the service (levels)
- John Dickey. Co-founder and CTO Talari Networks
(2) Programmable data planes: the P4 language and introductory use cases
- LJ Wobker. Barefoot Networks
(1) Cloud: Its all about the service (levels)
Actual service levels and visibility of those service levels are key to cloud adoption. It all starts with per packet performance. As enterprises continue to evaluate moving their key application to the cloud is the weak link in the service chain the network to the cloud itself? What service level is network infrastructure able to be provide to end users today and in the future? When the mouse click hangs… where do IT administrators look first? In this presentation, John Dickey will be providing an early glimpse of research conducted on international cloud based application performance contrasting traditional IPSec access methods with SD-WAN approaches. (2) Programmable data planes: the P4 language and introductory use cases
(2) Programmable data planes: The P4 language and introductory use cases
Large scale cloud networks often have specific requirements around packet manipulation and particularly around gathering information from within the network. This complicated by the requirement to gather information at speeds that are challenging or impossible in traditional CPU based mechanisms (such as SNMP or CLI). P4 is a language designed to describe packet forwarding planes and allow for developers and operators to capture “Inband telemetry” directly from the network devices and then export that information to consumer applications or processes. The presentation will cover the basic structure of the language itself as well as a discussion of the Inband Network Telemetry protocol and specification – both of which are also available via open source git repositories for experimentation and modification.