Home | Blog | Software | Reviews and Features | Forum | Help | Donate | About us
topbanner_forum
  *

avatar image

Welcome, Guest. Please login or register.
Did you miss your activation email?

Login with username, password and session length
  • December 07, 2016, 10:12:05 AM
  • Proudly celebrating 10 years online.
  • Donate now to become a lifetime supporting member of the site and get a non-expiring license key for all of our programs.
  • donate

Author Topic: The future of networking - resources for: SDN, OpenFlow and Mininet  (Read 2388 times)

40hz

  • Supporting Member
  • Joined in 2007
  • **
  • Posts: 11,768
    • View Profile
    • Donate to Member
The notion of virtualization has been with us for a long time now. Virtual machines have liberated computing environments from their underlying hardware - with obvious benefits realized industry-wide.

Now that same virtualization model and discipline is finally coming to our network infrastructure with the advent of SDN (software defined networking) and OpenFlow.

If you're going to be working as a network professional in the near future, an understanding of network virtualization will be as much an essential skill for an IT professional as an understanding of machine virtualization is today. So now is the time to start getting up to speed. :mrgreen:

The following Stamford seminar lecture by Scott Shenker gives an excellent overview of what SDN is, where it came from, and where it's going. Scott discusses the key players, the ideas behind SDN, its architecture, mistakes made along the way, lessons learned, and the design changes and conclusions that have since been drawn. At a little over an hour, this is one of the single best high-level introductions to this initiative, presented by somebody who is directly involved in bringing it to market. Although moderately technical in a few places, it's still very understandable for anyone who has even a tiny amount of home networking experience. Well worth a view - and highly recommended:



If a Stamford lecture is a little too formal for your taste (or you only have 25 or so minutes to spare) Eli the Computer Guy has a nice funky down-home introduction you may prefer to watch instead:



Either presentation is good. Watching both, however, will definitely give you everything you’ll need to be aware of when it comes to SDN.

***

Once you've got some SDN/OpenFlow basics under your belt you'll probably be anxious to put some of it to use - or at least play with it in a lab setting. But switches and networking hardware are awfully expensive, right? Well...luckily, there is a free virtual lab environment available for download called Mininet.

From the website:

Quote
Mininet creates a realistic virtual network, running real kernel, switch and application code, on a single machine (VM, cloud or native), in seconds, with a single command...

Because you can easily interact with your network using the Mininet CLI (and API), customize it, share it with others, or deploy it on real hardware, Mininet is useful for development, teaching, and research.

Mininet is also a great way to develop, share, and experiment with OpenFlow and Software-Defined Networking systems.

Full information, documentation, and instructional videos are available at the Mininet website. Mininet can be run on any operating system in Virtualbox or under Linux.

Want to see it in action? Watch a quick introductory video to Mininet below:



People with previous experience with the "terminal mode" found on virtually every professional router (Cisco, HP, etc.) should feel right at home in fairly short order.

SDN! Cool tool. VERY cool! :Thmbsup:
« Last Edit: July 11, 2014, 02:39:31 PM by 40hz »

Shades

  • Member
  • Joined in 2006
  • **
  • Posts: 2,099
    • View Profile
    • Donate to Member
Re: The future of networking - resources for: SDN, OpenFlow and Mininet
« Reply #1 on: July 19, 2014, 02:10:58 PM »
After seeing (just) the 'Eli'-video the concept is clear and very promising indeed.

That video didn't remove some of my doubts about the system.

Current networking functionality is quite rigid and if changes are needed, they need to be applied (by one or more able bodies) all over the network topology. A hassle management-wise to say the least. SDN will get rid of that by allowing changes to be done centrally and on-the-fly.

But in my (simple) way of thinking, does this not introduce a single point-of-failure? Where a virus or malware can take your complete network hostage? For example, by assigning all available networking capacity to a randomly chosen functionality (mail, HTTP, FTP, VOIP, etc.) every milli-second? That would severely affect every computer in your topology at once.   

Stoic Joker

  • Honorary Member
  • Joined in 2008
  • **
  • Posts: 6,296
    • View Profile
    • www.StoicJoker.com
    • Donate to Member
Re: The future of networking - resources for: SDN, OpenFlow and Mininet
« Reply #2 on: July 19, 2014, 11:24:08 PM »
But in my (simple) way of thinking, does this not introduce a single point-of-failure? Where a virus or malware can take your complete network hostage?

I haven't had time to watch the either video, but I'm inclined to say yes. Much like one bad stick of RAM can torch 20+ virtual servers... *Shrug* ...It's a risk/reward thing.

40hz

  • Supporting Member
  • Joined in 2007
  • **
  • Posts: 11,768
    • View Profile
    • Donate to Member
Re: The future of networking - resources for: SDN, OpenFlow and Mininet
« Reply #3 on: July 20, 2014, 08:36:30 PM »
But in my (simple) way of thinking, does this not introduce a single point-of-failure? Where a virus or malware can take your complete network hostage?

I haven't had time to watch the either video, but I'm inclined to say yes. Much like one bad stick of RAM can torch 20+ virtual servers... *Shrug* ...It's a risk/reward thing.

Precisely. The usual network security and redundancy caveats apply. That part of the equation will never change.

But...SDN is (to me) more like a front end/configuration API. There''s nothing to say it needs to be virtual. OpenStack can run on a single box in firmware or in a cloud setting. It's real advantage is that it provides a common grammar, syntax, and command set. One where a network can be easily configured and tested before committing anything to hardware. So as a research and prototyping tool it's a big plus even if the actual network gets implemented totally differently.

Again, it's not a "be all, end all" solution. But it's a good deal better than much of what we've had to work with up till now.