Network Design in the Post Crash 2008 World

The focus on separating user intent from infrastructure is an important new insight into how cloud environments should be run.  To drive forward this approach, Group-Based Policy is currently being developed for both OpenStack and OpenDaylight open source projects. Mike Cohen, Director of Product Management and Sandeep Agrawal Senior Marketing Manager, both of whom work at Cisco Systems join me to talk about Group-Based Policy or using intent to manage infrastructure.       

Duration: 21 minutes 17 seconds

Lippis Intro/Analysis @ : 00:10 sec

Question 1 @ 2:44 sec: Sandeep, What is GBP and what problem is trying to solve.

Question 2 @ 4:28 sec: Mike, What are the architectural components associated with GBP?

Question 3 @ 5:41 sec: Sandeep, How does GBP contribute to an open SDN approach?

Question 4 @ 6:48 sec: Sandeep,What does GBP mean for security and governance?

Question 5 @ 8:51 sec: Mike, Does GBP offer a means to automate network configuration and change management? 

Question 6 @ 10:23 sec: Mike, One part of GBP is that it gathers the state of the infrastructure.  How does it accomplish this, is there a state database that’s created?

Question 7 @ 12:20 sec: Mike, Does GBP assist in the creation of an application dependency map that can be shared across multiple IT teams, such as storage, networking, servers/virtualization and application development?

Question 8 @ 13:41 sec: Mike, How are the policies that are created configured within the products that make up the infrastructure?

Question 9 @ 15:09 sec: Mike, GBP is being adopted by OpenStack and OpenDayLight, but how will it be implemented by the vendor community within products?

Question 10 @ 18:03 sec: Mike, How does GBP compare to other efforts such as:  Congress and ONOS?

For the Lippis/Ixia evaluation at iSimCity, Brocade submitted the VDX™ 6740, which is Brocade’s flagship Top-of-Rack (ToR) switch that’s built with Brocade’s Virtual Cluster Switching (VCS™) technology, an Ethernet fabric innovation that addresses the unique requirements of highly virtualized, cloud data center environments.  The line-rate, low latency VDX™ 6740 fixed configuration switch is available in two models: the VDX™ 6740 with up to 64 1/10 GbE SFP+ ports, and the VDX™ 6740T with 48 1/10 GBASE-T ports and four 40 GbE SFP+ ports. We populated and tested the switch’s 10 and 40 GbE ports for performance and power consumption. In this podcast, Gautam Roy, Senior Product Marketing Manager at Brocade, joins me to review the VDX 6740’s test results.

Duration: 9 minutes 30 seconds

Lippis Intro/Analysis @ : 00:10 sec

Question 1 @ 2:38 sec:  We start with a description of the VDX 6740 and its use cases.  

Question 2 @ 3:25 sec:  VDX 6740 Latency and Throughput performance test results. 

Question 3 @ 4:30 sec:  VDX 6740 IP Multicast performance test results. 

Question 4 @ 5:07 sec:  VDX 6740 10 and 40 GbE congestion test results. 

Question 5 @ 6:15 sec:  VDX 6740 cloud computing performance test results. 

Question 6 @ 7:44 sec:  VDX 6740 power consumption test measurements.  

Question 7 @ 9:00 sec:  Closing

At the Open Networking Users Group meeting this past October, CIOs and Chief Network Architects from large corporations discussed their Software Defined Networking plans, obstacles, requirements and experiences.  One of the loud-and-clear messages was that business unit managers are demanding self-service IT provisioning.  For CIOs to deliver on this demand, networks need to be part of the auto provisioning process of compute and storage.  Networks are holding back CIOs’ ability to deploy new applications and business process automation since network configuration time is still measured in weeks rather than the minutes it takes to configure compute and storage.  Charles Ferland, IBM's SDN Global Sales & Business Development Manager joins me to talk about the impact that Software Defined Networking will have on Business Process from a CIO perspective. 

NEC has launched its fifth version of its popular ProgrammableFlow Software-Defined Network (SDN) controller just when OpenFlow 1.3 is starting to be implemented in popular switches.  The NEC ProgrammableFlow® Networking Suite, was the first commercially available SDN solution to leverage the OpenFlow protocol, enabling network virtualization and allowing enterprises, data centers and service providers to deploy, control, monitor and manage secure multi-tenant network infrastructure.  Don Clark, Director of Strategy and Business Development at NEC joins me to talk about NEC’s SDN investment and the latest version of its controller, which you all can learn more about by attending ONUG at JPMorgan in NYC on Oct 29^th and 30^th.

Podcast Duration: 11 minutes 9 seconds

Lippis Intro/Analysis @ : 00:10 sec

Question 1 @ 2:10 sec: Nick: Don, Lets start with a description of the NEC ProgrammableFlow SDN controller that is what is it, what’s new and what type of SDNs does it enable?  

Question 2 @ 3:34 sec: Nick: Don, OpenFlow 1.3 support was added to the controller, what new SDN designs with that enable?

Question 3 @ 6:05 sec: Nick: Don, I’d like to talk a bit more about your Top of Rack switches - which do the NEC ProgrammableFlow SDN controller support, all openflow 1.3 devices?

Question 4 @ 8:03 sec: Nick: Don, there are visualization features within the NEC controller, can you talk to that feature set? 

Question 5 @ 9:36 sec: Nick: Don, how do network architects and designers start to build SDNs with the NEC ProgrammableFlow controller?

There is a fundamental shift occurring in network engineering skill set requirements as open networking and Software-Defined Networking technologies continue their rapid deployments.  DevOps will have a larger role in network purchases.  New network designs are emerging thanks to virtual overlays, white box networking, Linus based network-programming tools, OpenFlow based pSwitches and vSwitches etc.  The CCIE skills set is being augmented with SDN skills and knowhow.  To keep competitive and relevant in the networking industry job market you need SDN skills.  ONUG Academy offers five tutorials taught by the experts who are deploying and developing open networking/SDN standards plus technology including:

T1: Understanding And Deploying Virtual Networks by: Srini Seetharaman, Technical Lead for SDN at Deutsche Telecom



T2: Getting Started with OpenFlow Deployments by: Brent Salisbury, University of Kentucky



T3: Integrating OpenFlow and OpenStack by: Rob Sherwood, Open Networking Foundation Chair of Architecture and Framework working group



T4: Understanding White Box Networking Architecture and Economics by: JR Rivers, co-Founder and CEO of Cumulus Networks



T5: Writing SDN Applications On Popular Controllers by: Matt Davy and Chuck Black, Tallac Networks

These tutorials and instructors were defined and hand picked by the ONUG board respectively who are sending their networking, virtualization and DevOps teams, so should you.  Here’s a 3-minute video on why you should enroll in ONUG Academy.    

Over the past two years NetScout has invested in Network Monitoring Switch technology, acquiring Simenna in 2011 and ONPATH in 2012.  This is the first in a two-part podcast series were Steven Shalita, Vice President of Marketing at NetScout discusses NetScout’s increased investments in the network monitoring switch business.  In this podcast we discuss network monitoring switch and fabric market attributes and dynamics.  In the second podcast Steve and I review network monitoring design options, service delivery and the implications that virtualization and Software-Defined Networking have on monitoring architectures.

This is the second of a two-part podcast series were Steven Shalita, Vice President of Marketing at NetScout discusses NetScout’s increased investments in the network monitoring switch business.  In this podcast we review network monitoring design options, service delivery and the implications that virtualization and Software-Defined Networking have on monitoring architectures.  The first podcast on network monitoring switch and fabric market dynamics can be found here.

To assure 100% throughput, that is no packet loss or network performance degradation takes place during an inter-chassis switch over, Cisco invested in its third generation of Virtual Switch System or VSS.  It introduced support for four Supervisors within a VSS, which it calls VS4O for short.  VS40 supports a VSS pair to be resilient to up to three simultaneous supervisor failures; providing deterministic and automated recovery from failures and increases network uptime tremendously in the campus backbone.  Sudeep Padiyar Senior Product Manager at Cisco Systems and Charles Okochu, Network Engineer at MarketAxess join me to discuss Cisco’s new additions to the Catalyst 6500 and the resiliency network design options it affords to Cisco customers.

Cisco introduced enhancements to its VXLAN implementation in the Nexus 1000V virtual switch that overcomes the requirement for IP Multicast.  In addition Cisco is integrating vPath to VXLAN for service insertion that stitches virtual services into VXLAN overlay tunnels. Network services might include virtual firewalls, application delivery controllers, WAN optimization, network monitors, etc.  A final VXLAN challenge that network teams are facing is how to integrate VXLAN into legacy networks and with existing physical networks, appliances and the enterprise WAN. We’ll hear if Cisco has any solutions to get around this issue as well.  Han Yang of Cisco joins me to discuss VXLAN, without multicast and with vPath, and how it can be used to deploy virtual network overlays.

As the need for visibility of applications flowing inside networks grows, network architecture itself is changing thanks to network plus server virtualization in data centers, and now Software-Defined Networking.  Also trends like BYOD, virtualization, and application mobility are bringing complexity and a new dynamism to today’s networks.  Larry Hart, Vice President, Ixia/Anue joins me to discuss how Software-Defined Networking is enhancing the hot network visualization market.