What is the difference between shared and switched ethernet?
Ethernet is a broadcast based protocol. In shared ethernet packets are flooded to all hosts on the segment whether the packet is intended for them or not. Switched ethernet works differently. If a switch does not know the location of a host, the packets are flooded to all ports, until the host is found. From then on, packets are switched to that host, and no other host sees those frames. However, packets that have a broadcast address for the destination will be sent to all hosts.
What is layer 3 switched ethernet?
Layer 3 switched is a contradiction of terminology if one understands the OSI model. In the OSI model, frame switching takes place at layer 2. Switching is very fast because it takes place in hardware. A switch contains a CAM table that maps a hardware address to a specific port on the switch. Routing occurs at layer 3. Instead of using the CAM table, routers rely on the ARP table which is a mapping of IP address to hardware address. It is usually a lot slower that switching because it has traditionally relied on software to make the routing decisions. Layer 3 switching combines both CAM table and the ARP table, by mapping an IP address to a port. A layer 3 switch will recieve a packet and figure out that it needs to be routed to another network or not. Depending on which routing protocol you are using, the decision is made and the packet routed. Now say another following packet is bound for the same destination address. The switch will know this because it first looks at the destination hardware address, so it relies on the layer 2 fabric to then switch the packet. The routing decision is only made once, then then the following packets are switched out the appropriate port.
What is a VLAN?
A VLAN or virtual lan is nothing more than a LAN that has been defined logically. It is important to remember that the network segment is not nessicarily defined by the phyisical layer (1). VLANs can be defined several ways. You can have a port based VLAN, a protocol based VLAN (only Nortel and Cisco), or a subnet based VLAN. We're going to discuss protocol based VLANs because that is what we are using at MU. We use three different types of protocol based VLANS IP, IPX, and AppleTalk. Each VLAN is assigned a unique number or name. Our IP VLANS are numbered 1xxx, IPX VLANS are 2xxx, and AppleTalk VLANs are 3xxx. An IP VLAN will not see IPX traffic, nor AppleTalk, it will only see IP. Though VLANs can logically span several switches, traffic within the VLAN is always switched at layer 2. Traffic between the VLANs is layer 3 switched, except for AppleTalk because it is not able to be layer 3 switched. We use the AppleTalk VLANs to trunk the traffic to a router that will forward or route the AppleTalk traffic.What is the spanning tree protocol?
Spanning tree is very important in a switched network. Since we have VLANs that span multiple switches, we run the risk of creating a layer 2 loop because we connect the switches at more than one point. Spanning tree is an algorithm that will run to determine if a layer 2 loop is in existance. If it does detect one, one of the ports that is involved in the loop will be blocked. Not only can spanning tree be a safety measure, you can use it to implement redundancy. For example, if you connect two switches together by two separate links, and those switches are in the same VLAN, spanning tree will automatically block one of the links. So you create an intentional loop. If your active link breaks or fails, the spanning tree algorithm will run again, and unblock the previously blocked link.