Current Trends in DC Networking - VXLAN Overview

The data center has two main struggles when it comes to networking.

  1. Loop free layer two by sacrificing half the links.
  2. VM or workload mobility anywhere at any time.

Smaller shops might be fine with giving up a few links, but extending a L2 domain is asking for trouble.  Larger shops need every link they can get and demand the flexibility to move a VM anywhere. So we are being forced down a road of L2 everywhere. How do we do it?

Our current model is broken, or better said, it's out dated. We need a technology that is adaptable to the movement of the application. This will take  some time...

Right now though, we have VMs needing to move within pods or zones or even DCs. Yup and application teams still require L2 domains between all this mess. So how do we make everyone happy?

Well... VXLAN

Its not the protocol we need but its the protocol we deserve.


ONUG Spring 2016

I just returned home from ONUG Spring 2016 in Mountain View CA. I was lucky enough to be invited by my buddies Steven and Tom of Tech Field Day. More and more people are talking about ONUG over and WOOT I was finally able to attend!

I'm blogging my thoughts on the event in hopes to encourage more people and companies to attend and get involved with the ONUG community. They are doing great work and its only going to get better!


Current Trends in DC Networking Series

Im kicking off a new series on various DC networking trends. All of these are technologies that are hot right now and I've been wanting to dig into deeper.

I'm not sure how this series will flow as Im jumping in and blogging as I go. The series will be structured from the ground up starting with the network infrastructure and progressing as we build out the DC.

Here is my base idea on organizing topics. As always, feed back and topic suggestions are welcome.

1: VXLAN Leaf/Spine (Arista)
2: VXLAN Leaf/Spine (Cumulus)
3: Automate Leaf/Spine Deployment w/ Ansible
4: Open vSwitch
5: CoreOS Deployment
6: CoreOS and Docker Containers
7: Automate CoreOS/Container deployment
8: Automate it all!

Id love to throw OpenStack into the mix but there is a lot there and it might need its own series. Im not sure we will see how this goes.

Alright, let the fun begin!!

First up will be VXLAN Leaf/Spine with Arista vEOS


Network Automation Training - Network to Code

I was fortunate enough to attend a networking automation class the other week, held by Network to Code and instructed by Jason Edelman (@jedelman8). As an introduction to automation for us networking grunts, I was very impressed with the class and learned a ton. So I thought I would share my experience and hopefully encourage others to venture down this road.

Who is Network to Code

Jason is a highly respected engineer with a CCIE and well over 12 years experience in the networking trenches. For the past 3 years he has focused his efforts on network automation and programming, and has been a leader in developing multiple tools to help us simpletons interact with networking hardware via programming tools. If you don’t know who Jason is then I highly suggest you follow him, as this dude is leading the charge in how we will do our jobs in the coming years.

With Jason’s passion to help others grow in network automation, Network to Code was born and now the company is teaching multiple classes as year across the globe, doing workshops, and providing services to clients helping them adopt automation technologies.


What do I do now?

Over the past year, every week I saw blog post and technology announcements pass by my purview. But I glanced at it and set it to the side so I could focus on my CCIE studies. Like a kid receiving the toy store catalog in the mail, I was checking off all the amazing technologies I wanted to dig deeper into and anticipation continued to grew. I wanted to spend time with these new technologies but instead I was stuck labbing OSPF forwarding address scenarios, and RIP/EIGRP/OSPF redistribution again....

Well, now is the day! Im free. No more CCIE to deal with and burn-out is starting to fade away. I can do what ever the hell I want and you cant stop me!!!

So what do I focus on first?


The CCIE Journey Ends

For my forth attempt I walked in feeling pretty good. But thats how I was for attempt three. I jumped straight in and cleared TS no problem. Had a couple tickets hold me up and one I couldn't fully solve. Diagnostics was stressful but I made it. I'm still not a fan of this section as 30 min could screw your day up pretty quick. And there is a ton of ground to cover in that short of a time frame.

Configuration was smooth until a large section of my lab would not come online. Id go into detail but NDA... This was not good and I knew right there it would cost me my test. I set it aside and finished the rest of the lab as quickly as possible. I spent well over an hour troubleshooting this section and still no dice. Rebuilt config, rebooted, I tried everything. Nothing... 3 minutes until the test ended and Im staring at my screen thinking "Dude, another test just slipped through your hands" So I accepted my loss, wr mem and clicked End Exam.


CCIE: Attempt #3

My day started off great. The troubleshooting section was up first and I quickly found my rhythm and started knocking off tickets. When I finished troubleshooting I still had a massive 4-pointer un-resolved, but it was the only ticket I didn't complete so I was happy. That will earn me a PASS!

Diagnostics (diag) was not straight forward this time. One ticket was simple and the second two were large. Very difficult to sift through all the configuration and topologies to solve them. Questions were not as straight forward as my previous attempts and were somewhat confusing. I worked on the last ticket until my 30 minutes expired. Diag left me nervous.

Configuration was familiar territory as well and went smooth. Nothing really tripped me up and everything was straight forward (for the CCIE). I botched one section which cost me a good hour to clean up but I was way ahead on time, so no big deal. With a little over an hour to spare I had finished all configuration sections with only one or two small sub-sections not working.

I ran through each section and verified functionality and double checked any show commands requested. I then spent a couple minutes working through those two issues but decided to leave them. I had already added up my points to be 85% and I didn't want to break something major this late in the game.

I quit my lab and left feeling good about the day. At the very worst I figured I got the dreaded pass/pass/pass FAIL (cut-score) results. I felt good overall.

Results came in around 9PM. FAIL...

I opened the results and almost couldn't believe what I was seeing. I passed troubleshooting but failed configuration. Failing configuration is an understatement. I bombed the hell out of configuration. Of the 5 sections I only scored above 50% in one, Network Services.

The longer I reviewed the results the more shock and confusion took over. I couldn't believe it. How could this be right? Are you sure this is my test? It wasn't until a day later I noticed I also failed diag. Just adding insult to injury now!

I have no clue what did this or how I could even complete that much and score this low. I didn't break any requirements and just like previous attempts everything was straight forward. Even the layer-2 section I scored below 45%. How is that possible and still have a fully functioning layer 3 and VPNs?

I'm really at a loss for what happened outside of maybe fat-fingering everything or devices crashing after I saved configuration. Killer is I scored so low that I don't have the option for a re-read.

This one stings a lot, but I need to focus on how to pick up the peaces and build a strategy for attempt #4. I don't know what to think right now.