Presentation on theme: "CS332 Chapter 21 Q and A CS332 Victor Norman Spring 2014."— Presentation transcript:
CS332 Chapter 21 Q and A CS332 Victor Norman Spring 2014
Diff. between MAC and IP address Q: Can you clarify the difference between a MAC and an IP address? A: A MAC address is a unique name of an interface in the local network, for that network’s flavor of layer 2. An IP address is a globally unique name for an interface. (An interface is where a computer connects to the network.)
IP address changes Q: Is the only time an IP address for a computer change at boot up? A: That is the most common time. But, I think an address can change when a DHCP address lease runs out and is not renewed. But, I’ve never seen that happen.
Classful address assignment? Q: You said in class that Calvin has a class B network, but the book says classes aren't used anymore. Does that mean that Calvin bought the class B before subnet/classless addressing was implemented? A: Every address is still officially in a class, but address blocks are not assigned to “companies” only in /8, /16, or /24 blocks anymore.
Subnet address assignment Q: Where is the subnet address assigned? (Does this mean “Where is the subnet mask assigned?”?) A: When a router’s interfaces are configured, the network administrator assigns IP addresses and subnet masks to the interfaces. (Or the IP address/subnet mask could come via DHCP.) A DHCP server is also configured to send out the subnet mask with its allocated IP addresses.
Classless address implementation Q: How was subnet/classless addressing implemented? A: “Companies” could always subnet their internal networks. The change was that now ISPs could give out non-A, B, or C blocks to companies. The only change that I know of was in routing protocols which used to depend on the network address to specify the mask. Now, the mask had to be sent explicitly.
Ugly address ranges Q: Is there any nice way to avoid ugly address ranges when using classless addressing as is the case with, e.g., a /28 network (top of page 356)? A: After a while, you won’t think of these as ugly. They are beautiful, but you have to learn to see them as beautiful… (No, no nice way to avoid them.)
VLANs’ IP prefixes Q: Do VLANs have separate IP prefixes? A: Yes! They are separate LANs, so separate networks, so have unique network parts.
Good question… Q: Are /30 and /29 the same IP address? If they aren’t, how would you know without the subnet mask? A: Neither are actually an address of a host – the host part in both is all 0s. But, if they read.9, then you wouldn’t be able to tell them apart without the subnet mask. But, that’s OK, because routing is done by network part, which includes the subnet mask.
Layer 3 Multicasting Q: What is the point of multicasting from IP if you have a network with vlan switches? Wouldn't you already have the connected hosts logically separated at layer 2? A: Layer 3 multicasting allows hosts to be grouped regardless of what physical LAN they are on. You could have computers in a multicast group that are in Kansas, Virginia, Tokyo, Reykjavik, and Abu Dhabi.
Loopback addresses Q: So how many loopback addresses are there? The book makes it seem like it doesn't matter what the last 24 bits are. (E.g would be equal to ). A: I’ve often wondered about this myself. I’ve only ever seen , but the mask makes it seem like you could have multiple ones. (And a google search seems to confirm that.)
Loopback interface Q: How does a loopback interface exactly work? A: If an application (layer 5) sends to it, layers 4 and 3 handle the packet normally. Layer 2 (the loopback device) turns the packet around and sends it back up to layer 3 to be processed.
Directed vs. Limited Bcast Q: Can you explain the difference between directed broadcast and limited broadcast? A: Directed bcast is the network part, followed by all 1s. Limited bcast is all 1s ( ). Directed bcast packet is routed to the network and then bcast. Limited bcast is broadcast on local network. (Routers don’t forward directed bcast packets anymore, I think.)
Limited Bcast at wrong layer? Q: Why is limited broadcast address implemented at the IP layer? Isn't this something that should be done at the Ethernet layer since it pertains to a single network that the host is directly connected too? A: Question: if you want to send to all hosts, what value are you going to put in the destination IP address field? What value will you provide to the socket?
Multi-homed devices Q: How common are multi-homed devices? A: Any proxy server is multi-homed. You also can put machines in a “DMZ” that are multi- homed. (Machines can ftp to/from it from the Internet and internally, but you cannot get to internal machines from outside.) There are probably other examples I don’t know about…