3213 DataLink+Network Flashcards
How does a network layer packet travel
In terms of Layer structure
Sender: Network layer takes segments from transport layer, encapsulates to packet and pass to data link
Intermediate Router: Packets routed through intermediate nodes/routers
Receiver: Receive packet from lower layers, extracts and pass to transport.
Routing Algorithm used for what
Determine path of packets
Internet protocol (IP) pros/cons
Pros: Best effort connectionless delivery
Keep routers simple
Cons: Unreliable, packets may be lost, out of order, duplicated
No throughput guarantee, no delay guarantee, no QoS guarantee
Higher layer handles issues
IP Addressing—Network Layer Address
Interface: Connection between host/router and physical link (aka. port)
IP address: Each interface must have a globally unique IP address.
router usually has multiple ip addresses for each interface. Host usually has 1.
IP Addressing prefix + suffix
Host ID 0 and 1
Prefix defines network
Suffix defines connection to the node
If Host ID all 0s: used to identify network address
If all 1’s, used to broadcast to all nodes.
IP address notation
4 bytes of IP address (4 octets)
eg. 00000000 00000000 00000000 00000000
Such that 11111111 represents 255 max
Classful Addressing
Divide entire IPv4 address into 5 classes labeled A,B,C,D,E. Each class has different size for different types of networks.
Draw diagram on cheatsheet !!
Subnet Addressing + interfacing
For easier management, large network may partition into smaller networks called subnets
Subnet can interface physically without intermediate router
Subnet: How does it work
Add another level of hierarchy to IP address
‘network prefix + subnet number + host number’
Only visible within organization
Subnet Mask: Usage + How to use
To find length of network prefix use subnet mask : w.x.y.z
Bits corresponding to network prefix are set to 1, and 0 otherwise.
Subnet range calculations
Given the added mid-fix of a subnet, we have n amounts of subnets, with n hosts in each. Say 11 bits for host, we can have 2^11 hosts, then MINUS 2 to save 0’s for network address, and 1’s for broadcast address
Creating subnet for # of hosts
Given x bits for Host ID, subtract 2^n to get amount of hosts needed for each subnet.
Now subtract n bits from Host bits, and add it to the Network bit range.
Apply mask to network ID, and the last few bits of 2^n represent starting point for subnet hosts.
Solutions to IPv4 Exhaustion
Classless InterDomain Routing (CIDR):
- short term sol. Reduce waste in address allocation.
Network Address Translation(NAT):
- Single machine/router with an ip address representing many computers behind it.
IPv6: 128 bit space
Classless interdomain routing
Abandon notion of classful addressing
Key concept: length of network id prefix can be any length.
pros: high flexibility and scalability
a.b.c.d/x, where x reps # of leftmost bits for network id
Network address translation
Private IP addresses for private internet
Private IP addresses not forwarded to internet, therefore different priv nets can re-use same private ip addresses.
Only 1 ip address is required from ISP to support whole private network to connect to internet.
1 IP for router, but specify port for hosts
How to get IP addresses
ICANN manages ip add spaces, and gives out blocks to ISP’s
ISP’s assign IP addys to host and router interfaces using Dynamic host config protocol
DHCP
Dynamic host config protocol:
Dynamically get address from a server, host uses DHCP to discover:
its own ip addy,
its network subnet mask,
ip address for DNS name server
ip address for first hop default router
IP fragmentation + reassembly
Identification: all frags of IP packet contain same iden value
Flags: unused, dont frag, more frag
Fragment offset: identifies location of frag in a packet. Where in cur pack this frag belongs, all frags multiple of 8 except last..
IP routing protocols
Internet routing: autonomous system
IntraAS routing:
