Network Topologies

This topic is to discuss the following lesson:

Hi guys.
I was not able to find lessons about that.

Can you explain the difference between a LAN SWITCH and a SAN SWITCH?

Thank you very much

Hello Giovanni

The first thing we should do is determine the difference between a LAN and a SAN. A LAN is already well defined within the Introduction to LANs lesson. A SAN is a storage area network, which is a network architecture that is used to provide access to consolidated data storage. This is often used within a datacenter to provide servers with direct high-speed access to a shared storage repository.

SANs can use conventional Layer 2 networking protocols such as Ethernet, but must often use a specialized Layer 3 protocol called Fiber Channel (FC) which was developed specifically for SANs. Unusual for network protocols, FC was actually developed by the International Committee for Information Technology Standards (INCITS) which is an ANSI accredited body.

A SAN switch is simply a switch that has FC ports.
.
I hope this has been helpful!

Laz

The lecture says the buildings are connected with full mesh to two core layer switches. But I am confused because the two core switches are not connected and there is no redundancy. Could you explain pls.

Hello Pau

Yes, you are correct. I believe that the point of the diagram was to show the various interconnections that can be made without actually showing every individual link. However, I will let Rene know to clarify and possibly to modify the image.

Thanks again for pointing this out!

Laz

HI Laz,

Wouldn`t be a good idea interconnect core 1 with core 2?

Also we have the option of one more link redundancy between cores and NY, but I guess it`s much more expensive ports in NY router than ports of core switches.

Let me know your opinion about this design

Thanks.

Victor Hugo

Hello Victor

Yes, your suggestion is excellent. This would allow traffic between buildings to avoid going through the New York router, offloading a lot of work from that device. It would also provide a level of redundancy, in case one of the core routers failed (if you enable something like load-balanced routing, HSRP, Stacking, or VSS.

You are correct again, that it would be beneficial to enable redundancy at the New York router, for both links, as well as in hardware. The NY router is a single point of failure that would cause a disconnection between multiple sites if it failed.

It’s good to keep in mind that in the real world, the best network design isn’t always strictly about what is best technologically. It also depends on things like cost, to whom the network belongs, as well as expected traffic patterns and usage. For example:

  • If the two buildings are not expected to have any traffic between them, then you can go without a link between the cores.
  • if the two buildings belong to two administratively disparate departments of the same company, it may not be according to policy to unify their networks.
  • Sometimes the cost, as you correctly mentioned, is a limiting factor in what you can actually do.

So network design also includes the idea of “do as much as you can with what you have”.

I hope this has been helpful!

Laz