Resilient Network Infrastructure for Virtualization

Network topology 101 for Virtual Infrastructures

I usually don’t like writing about obvious matters. Yes, fire is hot – night is dark and ice is cold. But in recent times I’ve witnessed some network topology designs (?), that made me frown.

I admit, that in some cases the situation is based on a lack of budget or just structures that have grown over years. I can understand that and it’s no shame. It’s my job to give advices and help to re-design.

No matter how many resources you have – if you use them without thinking, it will never be enough.

On the other hand there are environments who boast with high class components that have cost a fortune and which are organized in such an inefficient way that it almost hurts.

This article is not intended as a networking deep-dive. It’s a shallow 101 about network design that should be common knowledge. It’s a guide for the novice but I’d be happy to get responses by experts too.

The Basics

First let’s start with four simple networking requirements for Virtual Infrastructures.

  • redundancy
  • resiliency
  • bandwidth
  • latency

„Resilient Network Infrastructure for Virtualization“ weiterlesen

HA failover und STP

Eine Verfahrensweise, die ich mir bei Wartungen an ESX Clustern zu eigen gemacht habe, ist HA zu deaktivieren. Insbesondere dann, wenn es sich um Wartungen an der Netzinfrastruktur handelt. Warum eigentlich? HA ist inzwischen ein robuster Dienst und die Netzinfrastruktur sollte ohnehin redundant ausgelegt sein. Die Erfahrung zeigte allerdings, dass trotz aller Redundanz, Arbeiten an der Netzinfrastruktur immer wieder zu unerwünschten Failover Aktionen des HA führten. Dabei versucht z.B. ein Host (vergeblich), die laufenden VMs eines anderen zu starten. „HA failover und STP“ weiterlesen