Tuesday, May 22, 2018

Configuring OTV - Site Redundancy - Multiple AEDs

In this blogtorial we are going to expand on my previous blogtorial "Configuring OTV - OTV Configuration and Verification" and configure an extra AED on the West Site to achieve site redundancy.

Here is the topology.



Since the underlay network configurations are the same, please see my previous blogtorial for that part of the configuration.

Monday, May 21, 2018

Configuring OTV - Unicast Transport Mode - OTV Adjacency Server

In this blogtorial, we will be configuring OTV using unicast-only transport mode as opposed to using multicast to discover neighbors etc. In order to use unicast-only transport, we have to enable and configure OTV Adjacency Server feature. Please see previous blogtorial (Configuring OTV - OTV Configuration and Verification) if you want more information on configuring OTV using multicast.

This topology is the same as my previous blogtorial.


What is an OTV Adjacency Server?


"Overlay Transport Virtualization (OTV) provides support for nonmulticast-capable, unicast-only core networks through the OTV Adjacency Server feature. An edge device is configured as an adjacency server (primary or secondary). All other edge devices are configured with the IPv4 addresses of the primary and secondary adjacency servers, after which the edge devices communicate their reachability and capability information to the primary and secondary adjacency servers. Wide-Area Networking Configuration Guide: Overlay Transport Virtualization, Cisco IOS XE Release 3S 2 OTV Adjacency Server Restrictions for OTV Adjacency Server You can configure more than one adjacency server per VPN. An adjacency server can serve multiple VPNs. An adjacency server can also connect an OTV site to one or more VPNs." - Cisco.com

First let's get the IP and the transport layer or the underlay network configured.

We will start with the East site.

Wednesday, May 16, 2018

Configuring OTV - OTV Configuration and Verification

In this blogtorial, I will go through OTV configuration on virtualized routers on Eve-NG and go through a few verification commands. I will also touch on key OTV terminologies and design considerations.

If you need any assistance on how to get Eve-NG up and running on google compute, please see my previous blogtorial - Configuring Eve-NG on Google Compute Engine

Here is the topology. Our goal is to establish layer 2 connectivity between 5.5.5.5 (ESXi-West-1) and 5.5.5.10 (ESXi-East-1). One of the key advantages of using OTV versus other technologies such as VPLS is that spanning-tree domain will not be extended between the sites. Other benefits include ARP suppression, Unknown flooding suppression, Hardware acceleration, and multi path tunneling.



First thing we will configure is the transport network or the underlay. We will start with the devices in the East Site.

Monday, April 30, 2018

Configuring Eve-NG on Google Compute Engine

In this blogtorial I will walk-through on how to deploy Eve-NG on the cloud and more specifically on Google Compute Engine. I've tried Eve-NG on AWS and Azure but neither of them actually worked since nested virtualization isn't supported or is a very convoluted process (i.e Ravello).  Google Compute Engine was also the cheapest option at the time of this writing. 

Why deploy Eve-NG on the clould?


This whole blogtorial came about because I wanted to study advanced topics in the CCIE DC curriculum such as VXLAN with BGP EVPN. In order to virtualize the type of routers capable of running these advanced DC technologies it requires a lot of RAM and CPU. In addition, I didn't need these routers to be running all the time wasting power, so I needed the flexibility of powering these devices on and off. I decided to search google on how to virtualize the topology and much to my surprise, I could not find an article that encompassed all the steps. After reading about 15 different articles and stitching together all the information, I decided perhaps I should write an article that has everything you need to know (the whole nine yards!!) on how to get Eve-NG on the cloud (Google Compute Engine). 

If you follow this article from step 1 to step Z, you will end up with a fully functioning google instance running Eve-NG. 

First, register for a free google compute engine account and then log into the console (I believe you get $300 in free credits ... so enjoy!!). Once you log into the console, click on "Activate Google Cloud Shell". 


Once you activate it you should see something very similar.