English
A widely used alternative to the `flat addressing scheme` is the `hierarchical addressing scheme`. This addressing scheme builds upon the fact that networks usually contain much more hosts than routers. In this case, a first solution to reduce the size of the forwarding tables is to create a hierarchy of addresses. This is the solution chosen by the post office since postal addresses contain a country, sometimes a state or province, a city, a street and finally a street number. When an envelope is forwarded by a post office in a remote country, it only looks at the destination country, while a post office in the same province will look at the city information. Only the post office responsible for a given city will look at the street name and only the postman will use the street number. `Hierarchical addresses` provide a similar solution for network addresses. For example, the address of an Internet host attached to a campus network could contain in the high-order bits an identification of the Internet Service Provider (ISP) that serves the campus network. Then, a subsequent block of bits identifies the campus network which is one of the customers of the ISP. Finally, the low order bits of the address identify the host in the campus network.
This hierarchical allocation of addresses can be applied in any type of network. In practice, the allocation of the addresses must follow the network topology. Usually, this is achieved by dividing the addressing space in consecutive blocks and then allocating these blocks to different parts of the network. In a small network, the simplest solution is to allocate one block of addresses to each network node and assign the host addresses from the attached node.
In the above figure, assume that the network uses 16 bits addresses and that the prefix `01001010` has been assigned to the entire network. Since the network contains four routers, the network operator could assign one block of sixty-four addresses to each router. `R1` would use address `0100101000000000` while `A` could use address `0100101000000001`. `R2` could be assigned all addresses from `0100101001000000` to `0100101001111111`. `R4` could then use `0100101011000000` and assign `0100101011000001` to `B`. Other allocation schemes are possible. For example, `R3` could be allocated a larger block of addresses than `R2` and `R4` could use a sub-block from `R3` 's address block.
The main advantage of hierarchical addresses is that it is possible to significantly reduce the size of the forwarding tables. In many networks, the number of routers can be several orders of magnitude smaller than the number of hosts. A campus network may contain a dozen routers and thousands of hosts. The largest Internet Services Providers typically contain no more than a few tens of thousands of routers but still serve tens or hundreds of millions of hosts.
Despite their popularity, `hierarchical addresses` have some drawbacks. Their first drawback is that a lookup in the forwarding table is more complex than when using `flat addresses`. For example, on the Internet, network nodes have to perform a longest-match to forward each packet. This is partially compensated by the reduction in the size of the forwarding tables, but the additional complexity of the lookup operation has been a difficulty to implement hardware support for packet forwarding. A second drawback of the utilization of hierarchical addresses is that when a host connects for the first time to a network, it must contact one router to determine its own address. This requires some packet exchanges between the host and some routers. Furthermore, if a host moves and is attached to another routers, its network address will change. This can be an issue with some mobile hosts.
Dealing with heterogeneous datalink layers
Sometimes, the network layer needs to deal with heterogeneous datalink layers. For example, two hosts connected to different datalink layers exchange packets via routers that are using other types of datalink layers. Thanks to the network layer, this exchange of packets is possible provided that each packet can be placed inside a datalink layer frame before being transmitted. If all datalink layers support the same frame size, this is simple. When a node receives a frame, it decapsulates the packet that it contains, checks the header and forwards it, encapsulated inside another frame, to the outgoing interface. Unfortunately, the encapsulation operation is not always possible. Each datalink layer is characterized by the maximum frame size that it supports. Datalink layers typically support frames containing up to a few hundreds or a few thousands of bytes. The maximum frame size that a given datalink layer supports depends on its underlying technology. Unfortunately, most datalink layers support a different maximum frame size. This implies that when a host sends a large packet inside a frame to its nexthop router, there is a risk that this packet will have to traverse a link that is not capable of forwarding the packet inside a single frame. In principle, there are three possibilities to solve this problem. To discuss them, we consider a simple scenario with two hosts connected to a router as shown in the figure below.
Consider in the network above that host `A` wants to send a 900 bytes packet (870 bytes of payload and 30 bytes of header) to host `B` via router `R1`. Host `A` encapsulates this packet inside a single frame. The frame is received by router `R1` which extracts the packet. Router `R1` has three possible options to process this packet.
The packet is too large and router `R1` cannot forward it to router `R2`. It rejects the packet and sends a control packet back to the source (host `A`) to indicate that it cannot forward packets longer than 500 bytes (minus the packet header). The source could react to this control packet by retransmitting the information in smaller packets.
The network layer is able to fragment a packet. In our example, the router could fragment the packet in two parts. The first part contains the beginning of the payload and the second the end. There are two possible ways to perform this fragmentation.
Router `R1` fragments the packet into two fragments before transmitting them to router `R2`. Router `R2` reassembles the two packet fragments in a larger packet before transmitting them on the link towards host `B`.
Each of the packet fragments is a valid packet that contains a header with the source (host `A`) and destination (host `B`) addresses. When router `R2` receives a packet fragment, it treats this packet as a regular packet and forwards it to its final destination (host `B`). Host `B` reassembles the received fragments.
These three solutions have advantages and drawbacks. With the first solution, routers remain simple and do not need to perform any fragmentation operation. This is important when routers are implemented mainly in hardware. However, hosts must be complex since they need to store the packets that they produce if they need to pass through a link that does not support large packets. This increases the buffering required on the hosts.
Furthermore, a single large packet may potentially need to be retransmitted several times. Consider for example a network similar to the one shown above but with four routers. Assume that the link `R1->R2` supports 1000 bytes packets, link `R2->R3` 800 bytes packets and link `R3->R4` 600 bytes packets. A host attached to `R1` that sends large packet will have to first try 1000 bytes, then 800 bytes and finally 600 bytes. Fortunately, this scenario does not occur very often in practice and this is the reason why this solution is used in real networks.
Fragmenting packets on a per-link basis, as presented for the second solution, can minimize the transmission overhead since a packet is only fragmented on the links where fragmentation is required. Large packets can continue to be used downstream of a link that only accepts small packets. However, this reduction of the overhead comes with two drawbacks. First, fragmenting packets, potentially on all links, increases the processing time and the buffer requirements on the routers. Second, this solution leads to a longer end-to-end delay since the downstream router has to reassemble all the packet fragments before forwarding the packet.
The last solution is a compromise between the two others. Routers need to perform fragmentation but they do not need to reassemble packet fragments. Only the hosts need to have buffers to reassemble the received fragments. This solution has a lower end-to-end delay and requires fewer processing time and memory on the routers.
The first solution to the fragmentation problem presented above suggests the utilization of control packets to inform the source about the reception of a too long packet. This is only one of the functions that are performed by the control protocol in the network layer. Other functions include :
sending a control packet back to the source if a packet is received by a router that does not have a valid entry in its forwarding table
sending a control packet back to the source if a router detects that a packet is looping inside the network
verifying that packets can reach a given destination