Install and Configure Windows 2003 DHCP
The following tutorials run consecutively:
- Install DHCP Components
- Configure the First Scope
- Create DHCP Reservation
- DHCP Server Options
- The DHCP Process
- Troubleshooting DHCP
- DHCP Subnetting
- DHCP Supernetting
Open the Windows Components section of Add/Remove programs:
Open Network Services
Ensure that the DHCP checkbox is ticked and press OK
Wait
Press Finish
Using the 80/20 Rule for Servers and Scopes
To provide fault tolerance for the DHCP service within a given subnet, you can configure two DHCP servers to assign addresses on the same subnet. Here if one server fails then the other can take over. For balancing DHCP server use in this case, a good practice is to use the 80/20 rule to divide the scope addresses between the two DHCP servers. Server 1 is configured to assign 80% of the total addresses and server 2 is assigned the other 20%. Both servers have the same IP range but exclude each others portion of that range.
Configure the First Scope
Open the DHCP Management Console from Administrative Tools
The console with nothing configured, you should create a scope before you authorize the server.
Select New Scope
Press Next
Give the scope a relevant name, perhaps an indication of the location or purpose of the scope.
The IP Address range is important, dont use too many address in one scope when those extra addresses can be used somewhere else. If you only need one scope and you have less that 254 computers and network components then use the 192.168.x.y range with a Length of 24. With this configuration every separate subnet will increment the value of x. Example: 192.168.0.254 is the last IP address of the first subnet and 192.168.1.1 is the first IP address of the next subnet.
Once the IP range has been configured press Next
Here you can decide what IP address that are within the scope are not to be used for DHCP clients. The common standard as I know it is to use the first IP address for the Router or Default Gateway of the subnet. The next 10 or 20 IP Addresses are to be used for any Servers that you have on the subnet, although not necessary Domain Controllers, DNS servers, etc should use static IP addresses and hence these addresses should be excluded or not included within the DHCP scope. #
The lease duration is the length of time that a computer can use the IP address that it was assigned from the DHCP server. Unless you have a specific reason to change it just leave it as it is.
Select Yes and press next.
Add the IP address of the subnets router (Default Gateway). Dont use more than one.
Add the IP addresses of the DNS servers to be used by the DHCP clients. DNS servers do not have to be on the same subnet, they clients will attempt to contact the servers in the order that they are appear in this list. If you have a dns server on the local subnet then put that at the top.
Much the same as DNS servers but used for pre-windows 2000 name resolution. Do not configure if you have no Windows 9x/ME clients
Unless you have a reason not to, Select Yes and press next
Press Finish
The scope is shown and further configuration is possible from this window
Only Domain Controllers and Domain member servers can be authorized in Active Directory. Stand-alone DHCP servers or workgroup DHCP servers running windows 2000 or 2003 cannot be authorized in Active Directory but can coexist on the network as long as they are not deployed on the same subnet as an authorized DHCP server.
All working now.
Create DHCP Reservation
A DHCP Reservation is an IP address that is specifically designated to a particular Hardware (MAC) address. This IP address will not be leased to any other network component. Although the network component will always receive the same IP address, it will still be treated as a DHCP client and if the DHCP service is lost it will lose its address after the lease expires. A reservation is most commonly used for Network Printers.
Select New Reservation
The Reservation name should reflect the purpose of the lease or the location of the network component. The IP address must exist within the scope range. The MAC address is the 12 digit Hexadecimal Hardware address that if your lucky is usually written on the side of a printer. MAC addresses are unique. Description does not affect the operation of the reservation so write anything you like. BOOTP is a previous version of a Automatic Client Addressing System and is not necessary if all network components are DHCP compliant.
Further Reservation options that are independent of the scope options can be configured here. Reservation Complete.
DHCP Server Options
Unless you have a reason to, dont change anything.
Set options as above.
Again, Unless you have a reason to, dont change anything.
All Done.
The DHCP Process
Initial Lease Process
1 – The Client broadcasts a DHCP discover message to the local subnet.
2 – The DHCP server responds with a DHCP Offer message.
3 – If no response comes from a server the client can do one of two things:
- If the client is running windows 2000, it configures itself with an APIPA address (196.254.x.x)
- If the client is XP or 2003, the client configures itself with an alternate address, (if configured) or an APIPA address
4 – As soon as a DHCP Offer message is received, the client selects the offered address by replying to the server with a DHCP request message. Typically, the offering server sends a DHCP Acknowledgement (DHCP ACK) message approving the lease. (DHCP options are included in the acknowledgement)
5 – Once the client receives acknowledgment, it configured its TCP/IP properties using the info in the reply and joins the network.
DHCP Process
DHCP Client DHCP Server
-> -> DHCP Discover -> ->
<- <- DHCP Offer <- <-
-> -> DHCP Request -> ->
<- <- DHCP ACK <- <-
Lease Renewal Process
When a DHCP client restarts, it typically obtains a lease for the same IP address it had prior to restarting. Leases are renewed after 50% of the client lease time has elapsed, or when the IPConfig /renew command is initiated.
When the IPConfig /renew command is run the follows process occurs:
1 – The client sends a DHCP request message directly to the DHCP server that leased it, to renew and extend its current address lease.
2 – If the server is reachable, it will send a DHCP ACK to the client, which renews the lease and ends the session. If any config options have changed they will be updated in the DHCP ACK.
3 – If the DHCP server is unreachable the client waits until it is in the “Rebinding State”. By default this state occurs seven days after the initial lease started. When this state occurs it attempts to renew the lease from any DHCP server.
4 – If another server responds with a DHCP offer message to update the current client lease, the client can accept this offer and receive a new lease.
5 – If the lease expires and no server is available, the client must immediately discontinue using its leased IP address.
6 – The client then begins the initialization process.
In some cases the server can reply with a DHCP Negative Acknowledgment message to the client. The NACK message is sent to a client has requested cannot be provided by the DHCP server. This situation can occur when a client requests an invalid address or duplicate address for the network. If a client receives a negative acknowledgement, the lease renewal fails. The client begins a new lease initialization.
DHCP Messages
DHCP Discover
The destination address is shown as 255.255.255.255 which is the broadcast address. The source address is 0.0.0.0. The DHCP section identifies the packet as a Discover message. The client is identified by its MAC address.
DHCP Offer
The source address is now the server IP address and the destination is the broadcast address. The packet is identified as ana offer. The Your IP Address (Yiaddr) field contains the IP address that is being offered to the client. The DHCP Option field contains the various options being sent by the server. Options include the subnet mask, default gateway, lease time, WINS server address and NetBIOS info.
DHCP Request
The source IP of the client is still 0.0.0.0 and the destination for the packet is still 255.255.255.255. The client retains 0.0.0.0 because it hasn’t received approval from the server to start using the IP address that was given in the offer. The destination is still broadcast because more than one DHCP server may have responded. Broadcasting one particular requested address lets those other DHCP servers know that they can release their offered addresses and return them to their available pools. If the client has previously had a DHCP assigned IP address and the client is restarted, the client specifically requests this previously leased IP address in the DHCP Request field of the packet. If the server determines that the client can still use the address, it either remains silent or sends a DHCP ACK message. If the server determines that the client cannot have the address, it sends a NACK.
DHCP ACK
This message contains the IP address for the client to use as its lease. The source address of the ACK message is the server address, the destination address is still 255.255.255.255. The packet is identified as an ACK. The Your IP Address Yiaddr field contains the client’s address and the Client Ethernet Address Chaddr field contains the MAC address of the client’s network card.
DHCP NACK
A DHCP NACK address is most often used when the client computer has been moved to a new location. However, the message can also indicate that the client’s lease with the server has expired. All address fields are 0.0.0.0. After receiving the NACK, the client starts the DHCP Discover process again. The client attempts to lease the same address it had previously.
Troubleshooting DHCP
Address Conflicts
If a client has been assigned an address that is already in use, a warning will appear in the system tray. The system log will also shown address conflict info.
This conflict can be a sign of a DHCP scope error or rogue DHCP Server. The windows support tools includes dhcploc.exe which can be used to locate rogue DHCP servers. The Conflict Detection option from the DHCP server properties can be used to detect scope conflicts.
The shutdown /I command can be used to shutdown remote computers.
Using the Repair Button
Clicking the Repair button on the Support Tab of the Status dialog box performs the following actions:
1 – Broadcasts a DHCP request message to renew the current DHCP lease. Similar to the ipconfig /renew function except that ipconfig function sends the request by unicast – to the DHCP servers IP address - whereas the repair function uses a broadcast.
2 – Flush the ARP cache, similar to the arp –d * command.
3 – Flush the NetBIOS cache. Similar to the nbtstat –R command.
4 – Flush the DNS cache similar to the ipconfig /flushdns command.
5 – Register the client’s NetBIOS name and IP address with a WINS server, similar to the nbtstat –RR command.
6 – reregister the client’s computer name and IP address with DNS, similar to the ipconfig /registerdns command.
Failure to obtain a DHCP address
If the client has assigned itself an APIPA or alternate configuration address, the ipconfig /renew command or repair function can be used to correct the problem. If the problem exists then this indicates a problem connecting to the DHCP server or DHCP agent. Verifying the configuration of the DHCP server and agent.
The netsh dhcp show server command will show the names and addresses of all DHCP servers in Active Directory.
Verifying the TCP/IP installation and network hardware can be done by the ping localhost command, if this command returns a reply these two components are working. Pinging other hosts will verify network cables and switches etc.
Address Obtained from Incorrect Scope
DHCP request messages contain a field named Giaddr that informs the DHCP server of the originating subnet of the request. When the field is empty, the client is assigned an address from the local scope. When the Giaddr field contains an address the DHCP server will assign an address that is on the same subnet.
Verifying the Scope Configuration
First, verify that the scope is activated and make sure that the address range for the scope has been properly configured. For scopes that assign addresses for the server’s local subnet, ensure that the network id of the scope is the same as the local subnet, especially if the subnet id is not the usual /8 /16 /24. As an alternative, you can accommodate more computers within your current available address space simply by decreasing the lease duration in the scope properties. When the lease duration is shortened, computers that are shutdown, or removed from the network, do not keep their addresses for long and hence the address can be available for other hosts.
Next, check the exclusions for any static IP’s that might not be included. Move on to reservations and check that a reservation is not excluded, also check that the reserved address is within the scope. Also check the MAC address entry is correct. For network that use multiple DHCP servers, check that each server does not use address on the other servers.
Reconciling the DHCP Database
If you detect that DHCP database info is missing or inconsistent, you can attempt to resolve the problem by reconciling DHCP data for all or any scopes.
Scope IP lease info is stored in two forms by the DHCP server service:
1 – Detailed IP address lease info, stored in the DHCP database
2 – Summary IP address lease info, stored in the DHCP database
When the reconciling scopes, the detail and summary entries are compared to find inconsistencies.
In this process, the DHCP server either returns the addresses in question to their original owners or creates a temporary reservation for these addresses. These reservations are valid for the lease time assigned to the scope. When the lease time expires, the addresses are recovered for future use.
Subnetting
What is Subnetting
Subnetting refers to the practice of logically dividing a network address space by extending the string of 1 – bits used in the subnet mask of a network.
For Example, when the default subnet mask of 255.255.0.0 is used for hosts within the Class B network of 131.107.0.0, the IP address 131.107.0.1 and 131.107.255.254 are found on the same subnet, and these hosts communicate with each other by means of a broadcast. However if the subnet mask is extended to 255.255.255.0 the two IP address are then placed on different subnets. In order to communicate a default gateway must be used.
This process decreases the size of the broadcast domain, but reduces the number of hosts allowed per network subnet.
Subnetting also increases security by restricting network traffic behind routers.
Restricting Broadcast Traffic
A broadcast is a network message sent from a single host and distributed to all other network devices on the same physical network segment. Broadcasts use network bandwidth and every host that receives it must determine if it is destined for them and reply if necessary.
Routers block broadcasts.
Determining Host Capacity for Networks
For any specific network address, you can determine the quantity of host addresses available within that network by raising 2 to the number of bits used for the Host ID, and then subtracting 2. The network address 192.168.0.0./24 uses 8 bits for the host, therefore the number of hosts available is 2^8 – 2 = 254.
Excluding All 0’s and All 1’s Host ID’s
The value 2^x gives the total number of bit combinations for a binary number of x bits. 2^3 gives:
000 = 0
001 = 1
010 = 2
011 = 3
100 = 4
101 = 5
110 = 6
111 = 7
Not all combinations can be used, The All 0’s host ID is used for the local subnet. The All 1’s host ID is used for the broadcast address, hence when calculating the number of available hosts on a network you must subtract these 2.
Determining Subnet Capacity
When the string of 1’s in the subnet mask is extended beyond the default to create multiple subnets within an address space, The Host ID is shortened. And a new address space for the subnet ID’s is created.
To determine the number of subnets available within an address space, simply calculate the value of 2^y, where y is the number of bits in the Subnet ID. For Example, when the network address space 172.16.0.0/16 is subnetted to /24, 8 bits are left for the subnet ID. Therefore the number of available subnets is 2^8 = 256. you do not have to subtract 2 because modern routers including MS RRAS can accept subnets made up of all 1’s or 0’s.
Hosts per Subnet
The number of hosts available on a subnet is 2^x – 2. To calculate the number of hosts available to the entire subnetted network, simply multiply the number of hosts per subnet by the number of subnets.
Subnet Examples
The subnet mask does not need to be extended by a full octet. For the address space 10.0.0.0/12, the default subnet mask is 255.0.0.0 but it has been extended by 4 bits. Thus 4 bits have been borrowed from the Host ID and given to the Subnet ID.
Network ID Subnet ID Host ID
8 Bits 4 Bits 20 Bits
00001010 0000 0000 00000000
Number of subnets Number of Hosts per Subnet Total number of Hosts
2^4 = 32 2^20 – 2 = 1048574 32 * 1048574 = 33554368
The range of IP address available in the first subnet is 10.0.0.1 – 10.15.255.254
Estimating Subnet Address Ranges
You can estimate the IP range in each subnet by subtracting from 256 the value of the relevant octet in the subnet mask. For Example, for a Class C network such as 207.209.68.0 with a subnet mask of 255.255.255.192, subtracting 192 from 256 results in a value of 64. Hence the network’s subnet ranges are grouped in segments of 64 hosts. The first subnet range would then be 207.209.68.0 – 207.209.68.63.
For the Class B network 131.107.0.0 with a subnet mask of 255.255.240.0, subtracting 240 from 256 gives 16. Hence the subnets are grouped into segments of 16. Therefore the subnet addresses ranges reveal groupings of 16 in the third octet. The fourth octet still ranges as normal from 0 – 255, giving the first IP range as: 131.107.0.0 – 131.107.15.255.
Remember that hosts cannot be assigned an all 1’s or all 0’s address so those addresses of each subnet cannot be assigned.
Supernetting
Summarizing Routes Through Supernetting
To prevent depletion of higher-class network ID’s, the Internet Authority devised a scheme called Supernetting, which allows many networks (routes) to be grouped together (or summarized) in a single larger network.
For Example, suppose an organization needs to accommodate 2000 hosts. This number is too large for a Class C network ID. A Class B network can be used but there will be 63,534 unused addresses. Supernetting allows the organization to be assigned a block of Class C addresses that can be treated as a single network somewhere between a Class C and Class B address.
How Supernetting Works
Supernetting borrows bits from the network ID and masks them as part of the Host ID. Suppose you had a block of Class C addresses ranging from 207.46.168.0 to 207.46.175.0, by assigning a subnet mask of /21 instead of the default /24 to your routers and hosts results with your entire network being seen as a single network segment, because now that the Network ID has been shortened, the network ID of each host is now seen as being identical.
Below is a supernetted block of Class C addresses
Class C Networks Supernet ID Host ID
21 bits 11 bits
207.46.168.0 11001111 00101110 10101 000 00000000
207.46.169.0 11001111 00101110 10101 001 00000000
207.46.170.0 11001111 00101110 10101 010 00000000
207.46.171.0 11001111 00101110 10101 011 00000000
207.46.172.0 11001111 00101110 10101 100 00000000
207.46.173.0 11001111 00101110 10101 101 00000000
207.46.174.0 11001111 00101110 10101 110 00000000
207.46.175.0 11001111 00101110 10101 111 00000000
Subnet Mask
255.255.248.0 11111111 11111111 11111 000 00000000
Using Classless Interdomain Routing
CIDR is an efficient method of accounting for supernets within route tables. Were it not for CIDR, route tables would need a separate entry to handle every original network in the supernet. CIDR allows the supernet to be handled as a single entry.
CIDR is commonly used to refer to supernetting in general.
CIDR is not compatible with Routing Information Protocol RIP version 1. CIDR requires that routers use RIPv2 or Open Shortest Path First OSPF routing protocol.
Address Space Perspective
CIDR allows a block of Class C networks to be thought as a single address space in which 21 bits are fixed for the network ID and 11 bits are variable for the Host ID.
Using Variable – Length Subnet Masks
Traditionally, a singe subnet mask is shared by every host and router in an organization. This means that a network can be broken down into separate subnets. However, variable – length subnet masks (VSLMs), routers within an organization can handle different subnet masks. Most commonly, VSLM’s are used to allow subnets to be subnetted. For Example an organization with the address 131.107.0.0 / 16, means that routers external to the organization use the first 16 bits of the address to determine the network ID and route traffic appropriately. Once data has been received from the internet, the routers may use a subnet mask of /22 to route that data internally. Internal routers can again use a different subnet mask.
VSLM’s are not compatible with RIPv1.
Using VSLM’s to Accommodate Varying Subnet Sizes
Varying subnet sizes throughout the network means that IP addresses can be used more efficiently.
Class C Subnet Mask Options
Network Address Subnets Hosts per Subnet
208.147.66.0 / 24 1 254
208.147.66.0 / 25 2 126
208.147.66.0 / 26 4 62
208.147.66.0 / 27 8 30
When using VSLMs to divide your network into subnets of varying sizes, a specific pattern of subnet IDs with trailing 0s must be used, up to a maximum of seven subnets for a Class C network. These trailing 0s prevent the subnets from overlapping with each other.
No comments:
Post a Comment