bitmindframes

Grouping Devices into Networks and Hierarchical Addressing :
The original Internet – a few trusted users, no security issues.
Security needs have changed dramatically.
Dividing networks based on ownership means access can be prohibited, allowed, or monitored.
E.G. college network – admin, research, and student networks.
Security between networks is implemented in a router or firewall at the perimeter of the network.
The Internet = millions of hosts identified by a unique IP.
If each host had to know the address of all hosts it would severely degrade their performance.
Dividing networks so hosts who need to communicate are together reduces overhead (o/h) of hosts      needing to know all IPs.
For all other destinations, hosts only need to know the address of a router.
This router is called a gateway.  It serves as an exit from that network.
Hierarchical Addressing:
A hierarchical address uniquely identifies each host.
It also has levels that assist forwarding packets across internetworks, enabling division based on those levels.
So, hierarchical addressing supports communications between internetworks.
Postal addresses are examples of hierarchical addresses.
Refering only to the relevant address level (country, state, city, etc.) at each stage makes this process very efficient.
No need at each stage to know the all details (except Last).
IP addresses work in a similar way.
IP addresses have a network portion and a host portion.
Routers forward packets between networks using only the network portion.
Once delivered, the whole address will have been used to deliver the packet.
If necessary, additional layers of addressing can be created (subnet portion).

IPv4
Both network and host parts of IP are required.
For convenience IP addresses are divided in 4 octets.
Written in “dotted Decimal” notation for human readability.
EG. 192.168.18.57 (32 bits).
IP is hierarchical because routers only need to know how to reach the network, not the individual host.
EG. “.57” is the individual host.

The network portion is the same for all hosts in a network.
To divide a network the host portion can be cut into groups.
To do this we “borrow” bits from the host portion.
These borrowed host bits become network bits for the different subnets within the original network.
The more subnets created the fewer hosts per subnet.
Regardless, ALL 32 bits are required to identify individual hosts.
IPv4 Notation:
The # of bits in the network portion is called prefix length.
E.G. 192.168.18.57 -> network uses 24 bits so prefix = /24.
192.168.18.57/24
A separate 32-bit # (i.e. subnet mask) indicates the prefix.
255.255.255.0
Subnetting extends the prefix length or subnet mask.
Subnetting provides flexibility to divide networks.
In this chapter the examples all assume /24.
Fundamentals of Routes:
Within a network, hosts do not need a router.
But, to get “outside”, a router acts as a gateway.
Host configurations include a default gateway address.
This address = the router interface “closest” to the host.
Gateways Enable Communications between Networks.
To communicate with other networks, a host uses the default gateway to forward packets utside.
The router needs to know where to forward the packet next.
is called the next-hop address.
Routers forward packets to the next-hop that offers a path to the destination.
Packet Forwarding:
If the destination is in the same network as the source, a packet is delivered without the need for a router.
If not in the same network, a packet may be crossing many networks and routers. (payload is unaltered).
At each hop, forwarding decisions are based on information in the IP packet header.
Default Gateway:
Is needed to send a packet out of the local network.
If destination network does not equal source network, the packet sends to the gateway.
The Default Gateway (DG) network matches the network address of the hosts.
Host IP 192.168.1.23
Gateway IP 192.168.1.1
Packet Forwarding:
No packet can be forwarded without a route.
To forward the packets a host must have routes.
Routers make forwarding decisions for all packets (routing).
To forward packets a router requires a route to the destination
The destination network may be a number of hops away.
The route only indicates the next-hop, not the final router.
Routing:
Hosts route “outside” packets to the DG as their next hop.
PCs can add other manual routes, but this is unusual.
All routes on hosts or routers are stored in routing tables.
Routers automatically include routes for connected networks if configured with IP.
All other routes must be 1) configured manually, or 2) learned dynamically via a routing protocol.
Routing tables store information about directly connected networks and remote networks.
Routes in a routing table have three main features:
Destination network.
Next-hop address.
Metric (distance, bandwidth, etc.).
A router matches the destination address in a packet header with the destination network of a route in the routing table and forwards the packet to the next-hop.
If there are multiple routes, the metric is used to decide which route appears in the routing table (which is better in some way).
Packets cannot be forwarded without a route.
No route = dropped packet.
Routers may use a default route to forward packets when no specific routes are available.

View online info for a good example of Routing…..
“show ip route” to view on a router
Host Routing Table:
A host creates routes from its own IP configuration.
Hosts need a local routing table to direct packets correctly.
Unlike a router, the local table typically only contains its direct connections and its own default gateway.
netstat –r or route print will display host routing tables.

Routing Table Entries:
The route in a routing table, represents a range of host addresses (or even a range of network and host addresses).
Because IP is hierarchical, one route could refer to a large network and another to a subnet of that same network.
A router selects the most specific route.
The last and most specific address would be used.
However, if the street number was unknown, the third option would provide the best address match.
A packet to a subnet of a larger network would be routed using the route to the subnet.  A packet to another subnet within the larger network would use a more general entry.
E.G. for a packet with the destination of 10.1.1.55, the router uses the following precedence:
10.1.1.0 – most level of detail.
10.1.0.0 – next level of detail.
10.0.0.0 – next level of detail.
0.0.0.0 – (Default route if configured).
Dropped.
Default Route:
A route that will match all destination networks.
In IPv4 it is quad zero (0.0.0.0) or (0.0.0.0 0.0.0.0).
Packets that do not match more specific routes are forwarded to the next-hop router of the default route.
May be called the Gateway of Last Resort.
Next Hop Addresses and Packet Forwarding:
Next-hop = device that will process the packet next.
For a host, next hop = default gateway (router interface).
In a router, each route has a next hop for each destination.
As a packet arrives at a router, the destination address is compared to the routes in the routing table.
If a match is found, the packet is forwarded to that next hop.
The next-hop router is the gateway to networks beyond.
Networks directly connected to a router have no next-hop.
“R” = route
“C” = connected, no next-hop required
Instead, the router forwards packets directly to the host.

Packet Forwarding:
Routing is done packet-by-packet and hop-by-hop.
At each hop, the router examines the destination IP and checks the routing table for forwarding information.
The router will do 1 of 3 things with a packet:
Forward it to the next-hop router.
Forward it to the destination host.
Drop it.
Packet Examination:
Packets arrive at a router encapsulated as L2 PDU.
The router discards the L2 frame to examine the packet.
The destination IP address in the packet header is examined.
If a matching route is found -> forward packet out interface.
The packet has to be re-encapsulated at L2 and then forwarded out the interface.
The router removes the L2 encap.
Router extracts the dest IP add.
Router checks the routing table for a match

Traveling through an internetwork, a packet must find a route to its destination at every router.
If any router has no route (including a default route) the packet is dropped.
IP does NOT return a packet to the previous router.
ICMP will report error.
Routing
Each router does NOT need to know the path to all networks.
It only needs to know the next hop on the path.
The routing table = information used for forwarding decisions.
The routing table needs to have an accurate state of the network.
Out-of-date information can cause delays or packet loss.
Route information can be manually added to a router or learned dynamically from other routers.

Static Routing:
Refers to manually configuring routes on a router.
Default routes can also be statically configured.
Static routes are easy to configure for a few destinations.
Network changes -> manual updates required on every router!.
If not updated on time -> packet loss.
Dynamic Routing:
Maintaining routing tables manually is not always feasible.
Therefore, dynamic routing protocols are used.
Routing protocols = rules to dynamically share routing information. (RIP, EIGRP, OSPF, etc.)
As routers learn of changes they pass this to other routers.
When a router receives route updates, it updates its routing table and passes the information to other routers.
This way, all routers have accurate routing tables and they learn about routes that are many hops away.
Common routing protocols are:
RIP
EIGRP
OSPF
Routing protocols save time and work, but there are costs.
Exchanging route information consumes network bandwidth.
Routers must have processing capacity to implement the algorithms and to perform timely packet routing and forwarding.
Static routing consumes NO bandwidth or processing.
The cost for static routing is administrative.
In many internetworks, a combination of static, dynamic, and default routes are used to provide the necessary routes.

Tags: switch, interface, guides, vlan, ip, cisco