Engineers from La Salle-URL share the latest news and projects in the field of network solutions in telematic engineering.

15 April 2015 | Posted by Redacción Data Center

Networking requirements II

Hi, here we continue with our networking requirements posts. So far, by now we talk about the access network and SAN switches that Brocade can offer. But what about Cisco?

Cisco has a very large catalog of networking devices. For the access part we could use the well-known Catalyst 2960 switch. It is cheaper than other devices and very adaptable. For the SAN part we could use MDS 9000 Series Multi-layer switches.  They are ready to adapt to accommodate future mission-critical applications, massive amounts of data, and cloud environments. A fault-tolerant design and multi-protocol flexibility support non-stop operations and freedom of choice.

Now let's focus with some characteristics of our network. Some of this features are essential for the proper functioning of the network.

 

First of all we find redundancy.  In a network like ours we have to be careful. Our idea is to duplicate the switches and the number of links having at the end almost a full mesh topology. We said almost because we do not need to duplicate everything in the network because it will increase the price insanely.

Second we have load balancing. We do not want a part of the network to be overloaded. For this reason we would like to distribute and share the load among the network. That is why we will use an Brocade ADX 1000 or ADX 4000 which implements load balancing. Depending on the amount of data and cash we will choose one of those. 

Finally, and other important requirement is to avoid the Spanning Tree Protocol (STP). We don't want to use it because it takes at least 30 seconds to recover and this is too long in a data center. There are some ways to avoid it like VRRP(which is also a standard) with aggregation channels or GLBP which also implements load balancing.

Share