First let us define a flow: two packets belong to the same TCP (or UDP) flow if both packets carry a TCP segment (or a UDP datagram) that have the same source IP address, destination IP address, source port number, and destination port number.

For all the following questions, we assume that the NAT box has the following available ports for translation for TCP and UDP. When the NAT box needs to translate a packet for a new flow, it will pick the next available port number from the table below. We assume that the next available port number for TCP (resp., UDP) is at Index 1. If the port number at Index 1 is already used, the next available will be the one at Index 2 and so on. We assume that the NAT was just turned on. Therefore, the next available port # for translation of a TCP (resp. UDP) flow is 4006 (resp. 5007).

Index	TCP	UDP
1	4006	5007
2	6170	2750
3	1463	1070
4	1078	2553

As seen on Figure 5-55, a customer sets up a network with multiple machines. However, the Internet provider assigned only one IP address to the customer: this IP address is 128.194.65.02.

Consider a machine M on the customer's network "behind" a NAT box, i.e. on the premises of the customer (see Figure 5-55 above). The IP address of Machine M is 192.168.76.20. A TCP client socket is established at Port # 10100 to send an http query Q to the server at IP address 64.233.110.188.

Consider Packet P₁ carrying the http request Q. P₁ leaves Machine M and reaches Point A (see Figure 5-55). Fill in the array below with the source IP address, destination IP address, source port number, destination port number of Packet P₁ when it is at Point A. Tell whether this packet P₁ carries a TCP segment or a UDP datagram.

When the NAT box receives Packet P₁, it translates it and outputs on Point B (see Figure 5-55) a translated packet. Fill in the array below with the source IP address, destination IP address, source port number, destination port number of the translated Packet P₁. Highlight the information that will change (by the translation) and explain why.

When Packet P₁ reaches the server, the server will respond with a packet P₂. Fill in the array below with the source IP address, destination IP address, source port number, destination port number of the Packet P₂.

When the inbound packet P₂ reaches the NAT box, it will be translated. Fill in the array below with the source IP address, destination IP address, source port number, destination port number of the translated Packet P₂ (Point A).

Suppose now that the same machine M (IP address :192.168.76.20) makes a DNS request to the server at IP address 128.194.34.204. The DNS client is bound to Port # 10020. The DNS request will be carried by Packet P₃. Fill in the array below with the source IP address, destination IP address, source port number, destination port number of Packet P₃ (Point A).

When the NAT box receives Packet P₃, it translates it and outputs on Point B (see Figure 5-55) a translated packet. Fill in the array below with the source IP address, destination IP address, source port number, destination port number of the translated Packet P₃. Highlight the information that will change (by the translation) and explain why.

(to Internet) Figure 5-55. Placement and operation of a NAT box. OUTBOUND (to Internet) Figure 5-55. Placement and operation of a NAT box. OUTBOUND