Start a Conversation


This post is more than 5 years old


September 26th, 2012 07:00

SAN Exercise - Why Two-tiers of Switches

In Module 5's "FC SAN" exercise, the exercise solution had two tiers of switches.  The "edge switches" tier was connected to hosts, and the "core switches" tier was connected to arrays.  You had a total of 5 switches.

For a fabric of this size, why were two tiers necessary?  i.e. You could have had all of your hosts connected to 3 switches (as the video presented), with each of those switches connected to each array, couldn't you? 

For a modest sized SAN, with a limited number of arrays, why have two tiers of switches instead of a flat set of switches which connect to every array?


October 1st, 2012 06:00

Hello Kleinfelter,

I'm not sure of exactly which training course you are asking this question in regards to.   Can you please provide exactly which class this question is in regards to? 

As a generality you can create more flat SAN configurations with fewer switches but they aren't recommended by EMC.  For redundancy purposes you should wherever possible have multiple independent routes between the connections.   Many of the exercises only provide a limited number of hosts, storage, etc. listed in them just because they are that is needed to demonstrate the point.  In reality most production SAN configurations would be much more utilized with more connections.   Even in production environments careful attention would and should be used to build in redundancy for protection.

Thank you.

No Events found!