a couple of question regarding VLANs...first of all, why is there a need for them. Couldn't we achieve the same results by using dynamic multicast grouping using GMRP? Why design a separate protocol for this :?:
The second question is related to the VLAN tutorial on this site: In part 3, "Designing VLANS - A comparison with old netrworks", the diagram of the (bad) solution without using VLANs shows a bridge for every network and every level. If instead of this, we used only 3 bridges, one for every network, with each level connected via a separate port on the bridge, we will have reduced the number of bridges by 6! This would satisfy the low-cost arguement in favour of VLANs. Basically my point is that in this case we could have solved the problem using LAN segments.
Thanks for bearing with me. Awaiting your responses
Re: Why use VLANs? there are other ways...
12 years 8 months ago #10991
GMRP does not have the same effect as using VLANs. This protocol is used for identifying hosts that all want to receive the same thing from the same host...multicasting. It works along side GARP, the layer 2 mechanism for sharing this information, and IGMP, a layer 3 protocol for multicasting. VLANs, in a nutshell, place hosts into a configurable broadcast domain. Each individual host has their connections independent of each other, while in multicasting, the hosts are essentially sharing a common connection. Also, GMRP constrains multicasts to hosts within a common VLAN.
And as far as your second question, their illustration is to show the "old design" has the same benefits as using VLANs as far as security, but you can see the increased use of switches and unnecessary cabling, making VLANs superior. Your solution would make the network less secure by letting separate networks mix without the security mechanisms of VLANs.