Multitech RF500S Instrukcja Użytkownika Pobierz pdf (Strona 29)

foolproof since IP addresses can be spoofed, so caution is still advised. Both solutions eliminate the need to

reconfigure anything during fail over.

7.3 LAN Address Assignment

Each device on the network requires an address. The LAN uses private addresses. These addresses are not

used on the Internet therefore they do not need to be coordinated with other Internet users. They still must

be coordinated within your network since addresses cannot overlap.

7.3.1 Dynamic

In most cases dynamic address assignment is convenient. When a new machine is plugged in the DHCP

server, built into the router, assigns it an address. Once the device has an address is can use the LAN. The

DHCP server assigns several other critical numbers, a subnet mask and a gateway address. As discussed

previously the subnet mask defines how the address should be interpreted. Only machines on the same

subnet can directly communicate. The gateway address is where the computer sends any packets that

cannot be delivered locally. The gateway is the router. It is up to the router to determine how to deliver the

packet. In our case this is very simple, any message not addressed to a host on the LAN is forwarded to the

ISP. The router also tells the computer on the LAN what address to use for DNS lookup.

7.3.2 Static

For some devices, such as servers, dynamic addresses are a problem. The MultiTech router does not

implement DNS for LAN based devices. That means local servers need to be accessed by IP address rather

then name. We need the ability to permanently assign an address to a specific device. The router has a neat

solution for this problem. All addresses are dynamically assigned. However once an address is issued it can

be locked so it never changes. This is ideal each device thinks it is getting a dynamic address while the

administrator is able to freeze specific address.

The router performs this bit of magic by binding an IP address to the Ethernet MAC address. Each Ethernet

device has a unique 48-bit Media Access Controller address. This is how Ethernet devices communicate

with one another. This is much more convenient then setting IP addresses manually and making sure they

do not conflict with previously assigned addresses or the DHCP pool.

7.4 NAT -- Sharing a Single Internet Connection

The LAN cannot simply be “plugged in” to the Internet. Because the addresses used on the LAN cannot be

used on the Internet and because the ISP only provides a single address we need a translation mechanism

between the two networks. Network Address Translation (NAT) provides a mechanism to translate

addresses on one side to addresses on the other. When NAT is combined with private IP addresses we have

the ability to create a LAN with an unlimited number of local addresses and map them to a single public

address.

The way NAT works is that all addresses used on the LAN come from the private address pool. IntraLAN

communication proceeds normally NAT is not required. When a request cannot be serviced locally it is

passed to the NAT router. The router modifies the address and port number to match the public address

issued by the ISP and sends it on it way. When the reply comes back the router converts the address to that

of the original device and forwards it to the LAN. The NAT router can keep track of a large number of

sessions so multiple devices can use the same address.

For more information see RFC1631 The IP Network Address Translator (NAT).

1 2 ... 24 25 26 27 28 29 30 31 32 33 34 ... 48 49

Brak uwag

Multitech RF500S Instrukcja Użytkownika Strona 29