Network Fundamentals: VLANs, Spanning Tree, and Network Layer Concepts
VLANs and Hubs
VLANs require configuration tables in bridges to function. However, if hubs are used instead of switches, configuration tables are not necessary because hubs simply connect incoming lines electrically without performing any routing or configuration.
Spanning Tree Algorithm
Steps:
- Elect a root node (switch with the lowest address).
- Grow the tree as the shortest distances from the root (using the lowest address to break distance ties).
- Turn off ports for forwarding if they are not on the spanning tree.
Need for Network Layer
The design of networks with links, switches, and frames does not scale to large networks due to:
- Potential for routing table explosion
- Broadcast traffic causing network congestion
- Lack of support for multiple link layer technologies
- Limited traffic control using spanning tree
Distance Vector Routing Algorithm
Steps:
- Each node knows the distance of links to its neighbors.
- Each node advertises a vector of lowest known distances to all neighbors.
- Each node uses received vectors to update its own.
- Repeat periodically.
Connection-Oriented vs. Connectionless Service
Connection-Oriented:
- File transfer
- Remote login
- Video on demand
Connectionless:
- Credit card verification
- Electronic funds transfer
Bandwidth Requirements for Distance Vector Routing
With 8-bit costs and 50 routers, the routing table is 400 bits. Exchanging distance vectors twice a second requires 800 bps on each line in each direction.
Differentiated Services and Expedited Forwarding
There is no guarantee that expedited packets will experience shorter delays than regular packets. If too many packets are expedited, their channel may have worse performance than the regular channel.
IP Datagram Fragmentation and Strict Source Routing
The strict source routing option must be copied into each fragment because the information is needed to route every fragment.
IP Address Conversion and Subnet Mask
a) The IP address C22F1582 in hexadecimal notation is 194.47.21.130 in dotted decimal notation.
p>b) A subnet mask of 255.255.240.0 indicates a 20-bit network part, leaving 12 bits for hosts. Therefore, the maximum number of hosts is 2^12 = 4096.
Ethernet Addresses and IP Addresses
Ethernet addresses are not tied to specific networks because each Ethernet adapter is sold with a hardwired address that is not dependent on the location of the card.
IP Address Aggregation
The IP addresses 57.6.96.0/21, 57.6.104.0/21, 57.6.112.0/21, and 57.6.120.0/21 can be aggregated to 57.6.96.0/19.
IP Address Aggregation with Gaps
To handle a gap of 1024 unassigned addresses in the aggregated block 29.18.0.0/17, a new table entry can be added: 29.18.60.0/22. This allows for the assignment of large blocks to outgoing lines.
IP Fragment Reassembly
IP fragments may arrive out of order or be missing. To reassemble them, all pieces must be buffered until the last fragment arrives and the total size is known. A buffer of the right size is then built, and the fragments are placed into the buffer using a bit map to track which 8-byte blocks have been received.
IP Header Checksum
The IP header checksum covers only the header and not the data because many hosts do not check if a packet is delivered to them, assuming the network will not deliver packets from other hosts. Additionally, data checksumming can be expensive, and upper-layer protocols may already perform checksumming.
IPv6 Features
a) IPv6 uses 16-byte addresses, providing 3.4 × 10^38 addresses.
b) The Protocol field in the IPv4 header is not present in the IPv6 header because intermediate routers do not need to know which protocol handler to give the IP packet.
c) The ARP protocol does not need to be changed when IPv6 is introduced, but the requested IP addresses will be larger, requiring larger fields.