Determine the path between two hosts across a network 3

Asymmetric Routing

Because routers do not necessarily have the same information in their routing tables, packets can traverse the network in one direction, using one path, and return via another path. This is called asymmetric routing. Asymmetric routing is more common in the Internet, which uses the BGP routing protocol than it is in most internal networks.

This example implies that when designing and troubleshooting a network, the network administrator should check the following routing information:

Is there a path from source to destination available in both directions?

Is the path taken in both directions the same path? (Asymmetrical routing is not uncommon, but sometimes can pose additional issues.)

Packet Fields and Frame Fields

As we discussed previously, routers make their primary forwarding decision by examining the destination IP address of a packet. Before sending a packet out the proper exit interface, the IP packet needs to be encapsulated into a Layer 2 data link frame. Later in this section we will follow an IP packet from source to destination, examining the encapsulation and de-capsulation process at each router. But first, we will review the format of a Layer 3 IP packet and a Layer 2 Ethernet frame.

Internet Protocol (IP) Packet Format

The Internet Protocol specified in RFC 791 defines the IP packet format. The IP packet header has specific fields that contain information about the packet and about the sending and receiving hosts. Below is a list of the fields in the IP header and a brief description for each one. You should already be familiar with destination IP address, source IP address, version, and Time To Live (TTL) fields. The other fields are important but are outside the scope of this course.



Version – Version number (4 bits); predominant version is IP version 4 (IPv4)
IP header length – Header length in 32-bit words (4 bits)
Precedence and type of service – How the datagram should be handled (8 bits); the first 3 bits are precedence bits (this use has been superseded by Differentiated Services Code Point [DSCP], which uses the first 6 bits [last 2 reserved])
Packet length – Total length (header + data) (16 bits)
Identification - Unique IP datagram value (16 bits)
Flags – Controls fragmenting (3 bits)
Fragment offset – Supports fragmentation of datagram’s to allow differing maximum transmission units (MTUs) in the Internet (13 bits)
Time to Live (TTL) – Identifies how many routers can be traversed by the datagram before being dropped (8 bits)
Protocol – Upper-layer protocol sending the datagram (8 bits)
Header checksum – Integrity check on the header (16 bits)
Source IP address – 32-bit source IP address (32 bits)
Destination IP address – 32-bit destination IP address (32 bits)
IP options – Network testing, debugging, security, and others (0 or 32 bits, if any)