Transcript here

IST 220 Lectures for:
Dec. 8, 2009
Dec. 10, 2009
1
Introduction
• Each WAN or LAN belongs to a particular
organization, an internetwork does not belong to
any particular org
• An internetwork could actually be very small,
though it is usually big
• For big internetworks, the routing tables must be
still small; otherwise, too slow
• To make routing tables small, we would need to
use hierarchical addresses
– IP addresses are a hierarchical address scheme
• Each IP address has 32 bits
2
Five Classes of IP addresses
To make both big and small companies happy,
one class of IP address is NOT enough.
3
How to identify the class of an IP
address?
The first 4 bits of an IP address identifies its class.
0 * * * : class A IP address
1 0 * * : class B IP address
1 1 0 * : class C IP address
“*” means either
bit “1”
or bit “0”
1 1 1 0 : class D IP address
1 1 1 1 : class E IP address
4
Dotted Decimal Notation of an IP
Address
Remembering 32 bits IP addresses is often
too HARD for human being.
To solve this problem, we translate the 32 bits to decimal.
10000001
00110100
00000110
00000000
Binary
 Decimal:
129
.
52
.
6
.
0
5
How to know the Class of a
decimal IP address?
Example: 129.52.6.0
Step 1: translate the first decimal number “129”
to binary, we get: 10000001
Step 2: because the first 4 bits “1000” match
the identifier of Class A, it is a Class A address.
Idea: the first decimal number tells the Class info:
Suffix:
Class
1st decimal number
A
0 through 127
Last 3 decimal numbers
B
128-191
Last 2 decimal numbers
C
192-223
6
Last 1 decimal number
IP Packet Forwarding in a tiny
Internetwork
Question 1: how many physical networks are there?
Principle: each physical network belongs to a particular
organization; physical networks are interconnected
by routers, not bridges.
7
Answer to Question1
Answer: the two routers R1 and R2 break the
internetwork into 3 physical networks.
Note that Physical Network 2 includes two LAN
segments and the bridge.
Note that half of R1 belongs to Phy Network 1; the
other half of R1 belongs to Phy Network 3.
8
What is the Network Address of
Phy Network 1?
Each Phy Network has a unique network address
which cannot be used by any computer in it.
The “prefix” of the Network Address must be the same
as the prefix of the IP addresses of all other IP addresses
used in the network; but all the “suffix” bits must be “0”
Because the first decimal number of A’s IP address is
166, it is a Class B IP address.
The “prefix” of A’s IP address should be 166.200 because
the first 16 bits are the prefix of any Class B address.
9
What is the Network Address of
Phy Network 1 (cont.)?
Because all the IP addresses used in a Phy Network must
have the same prefix, the prefix of the IP addresses of all
Other computers, though now shown, must be 166.200
Hence, the “prefix” of the Network Address must be
166.200.
Hence, the network address is: 166.200.0.0
10
What is the Broadcast Address of
Phy Network 1 (cont.)?
The “prefix” of the Broadcast Address must be the same
as the prefix of the IP addresses of all other IP addresses
used in the network; but all the “suffix” bits must be “1”
Hence, the “prefix” of the Broadcast Address must be
166.200; and the suffix must be: 11111111-11111111
When we translate the 16 bits suffix back to
decimal, we will get: 255.255
Hence, the broadcast address is: 166.200.255.255
11
Maximally, Phy Network 1 can hold
how many computers?
Because it is a Class B network, the suffix part of the IP
addresses used in the Phy Network will have 16 bits.
Because the total number of all possible combinations
of the 16 bits is 216, the Phy Network can theoretically
hold 216 computers.
However, because neither the Network Address nor the
Broadcast Address can be used by a computer, maximally
the Phy Network can only hold 216 – 2 computers.
12
How many IP addresses should R1
own?
NIC1,1 belongs to Phy Network 1; however,
NIC 1,2 belongs to Phy Network 3.
Different Phy Networks must have diff. prefix.
Hence, the IP addresses assigned to these two
NIC cards must have different prefix.
Therefore, R1 owns two IP addresses.
13
Assume computer A sends
message “Hello” to D:
Step 1: P-A (processor of A) adds the “IP header” (also
called the routing layer header):
| 168.16.24.115 | 166.200.122.1 | “hello” |
(The IP layer payload includes the message only.)
Step 2: P-1 adds the NIC Layer (also called “Network
Interface Layer”) header:
| NIC-11’s MAC | NIC-A’s MAC | 168.16.24.115 | 166.200.122.1 | “hello” |
(The NIC Layer payload includes the message and the two IP addr.)
14
Assume computer A sends
message “Hello” to D (cont’d):
Step 3: NIC-A adds preamble and CRC
| Preamble | NIC-11’s MAC | NIC-A’s MAC |
168.16.24.115 | 166.200.122.1 | “hello” | CRC|
Step 4: NIC-A uses CSMA-CD to send the packet out after
modulation.
15
Assume computer A sends
message “Hello” to D (cont’d):
Step 5: NIC-11 senses the signal, demodulates it,
and puts it into cache
Step 6: NIC-11 does address filtering
Step 7: NIC-11 does length checking
Step 8: NIC-11 does CRC checking
Step 9: NIC-11 strips off preamble & CRC
Step 10: NIC-11 puts the packet into incoming queue
16
Assume computer A sends
message “Hello” to D (cont’d):
Step 11: P-1 (processor of R1) gets the packet from queue
Step 12: P-1 strips off the NIC layer header
Step 13: P-1 does routing based on the Routing Table.
Destination
166.200.0.0
167.234.0.0
168.16.0.0
Where to forward
NIC1,1
NIC1,2
NIC1,2
17
Assume computer A sends
message “Hello” to D (cont’d):
Notes on the Routing Table:
-The Destination column should contain Network Addr.
-The Dest field tells the receiver’s Phy Network
-The Routing Table should forward the packet towards
the shortest path, if there exists multiple paths from
R1 to the receiver.
-When P-1 matches the packet’s Dest IP Address against
the routing table, the last row will be matched because
the Network Address in the last row and the packet’s dest
IP address have the SAME prefix.
18
Assume computer A sends
message “Hello” to D (cont’d):
Step 14: P-1 adds a new NIC Layer header:
| NIC-21’s MAC | NIC-12’s MAC | 168.16.24.115 | 166.200.122.1 | “hello” |
Step 15: NIC-12 adds the preamble and CRC; NIC-12 then
uses CSMA-CD to send it out after modulation.
Step 16: NIC-21 will do exactly the same things NIC-11 has
done.
19
Assume computer A sends
message “Hello” to D (cont’d):
Step 17: P-2 will do exactly the same things P-1 has
done except that it will use a different routing table.
Destination
166.200.0.0
167.234.0.0
168.16.0.0
Where to forward
NIC2,1
NIC2,2
NIC2,2
20
Assume computer A sends
message “Hello” to D (cont’d):
Step 18: P-2 adds a new NIC layer header:
| NIC-D’s MAC | NIC-22’s MAC | 168.16.24.115 | 166.200.122.1 | “hello” |
Alert: Inside the NIC layer header, the dest address
cannot be the MAC address of NIC-L of the Bridge,
since no bridge does address filtering.
21
Assume computer A sends
message “Hello” to D (cont’d):
Step 19: NIC-22 adds the preamble and CRC; it then
uses CSMA-CD to send packet out after modulation.
Step 20: NIC-L will sense the signal, demodulates it,
and puts it into cache.
Step 20: NIC-L will do length checking;
Step 21: NIC-L will do CRC checking
Step 22: NIC-L will strip off preamble & CRC
Step 23: NIC-L will put the packet into queue.
22
Assume computer A sends
message “Hello” to D (cont’d):
Step 24: P-B (bridge) will decide where to forward the
packet based on its forwarding table:
Destination
Where is the
packet from?
Where to
forward?
D’s MAC Addr.
Left segment
NIC-R
D’s MAC Addr.
Right segment
discard
C’s MAC Addr.
Left segment
Discard
C’s MAC Addr.
Right segment
NIC-L
23
Assume computer A sends
message “Hello” to D (cont’d):
Step 25: because the first rule is matched, P-B will
forwards the packet to NIC-R.
Step 26: NIC-R adds preamble and CRC.
Step 27: NIC-R uses CSMA-CD to sent the packet out
after modulation.
Step 28: NIC-D senses the signal, demodulates it, and
puts it into cache.
Step 29: NIC-D does address filtering.
24
Assume computer A sends
message “Hello” to D (cont’d):
Step 30: NIC-D does length checking.
Step 31: NIC-D does CRC checking.
Step 32: NIC-D strips off preamble and CRC.
Step 33: NIC-D forwards the packet to P-C.
Step 34: P-D strips off the NIC layer header;
Step 35: P-D strips off the IP header after knowing who
the sender is.
Step 36: P-D shows the message in Outlook.
25