How to Create a Wired Network

Last updated: 25 Apr 2008

Overview

Overview of ethernet network and TCP/IP basics; selection of equipment and cables; hands-on experience creating a small LAN; connecting to the Internet; security concerns; testing network


Workshop Duration
2 hrs.

Student Prerequisites
none

Student Provides
  1. nothing

Lab Provides
  1. Two Computers
  2. Two disk drives with Windows XP and WireShark on it
  3. Administrator account and password on each
  4. Two wired ethernet NICs and cables
  5. One ethernet hub

Preparation
  1. Fresh installation of Windows XP SP2 with WireShark cloned to x disk drives
  2. Disk drives inserted and locked into computers
  3. Setup itadmin account with administrator rights and password
  4. Install Ethernet PCI Adapter
  5. Plug in ethernet hub
  6. Cable examples (internal twisted pair wires, crossover)

Delivery
  1. References
    • Ethernet: The Definitive Guide. Spurgeon, Charles E. O'Reilly & Associates. 2000.
    • TCP/IP Network Administration. Hunt, Craig. O'Reilly & Associates. 1994.
  2. Network Fundamentals
    • social, telephone, data networks
    • network model

      Seven-layer OSI model

      1. Layer 7: Application

        mechanisms to support end-user apps such as mail, file transfer, and web access

      2. Layer 6: Presentation

        mechanisms for data representation within applications

      3. Layer 5: Session

        mechanisms for reliable communications among cooperating applications

      4. Layer 4: Transport

        mechanisms for reliable end-to-end error recovery and flow control

      5. Layer 3: Network

        staion to station communication across a number of links -- an internetwork or internet.

      6. Layer 2: Data Link

        staion to station communication across one link -- transmit and receive frames/packets, recognize links, etc.

      7. Layer 1: Physical

        electrical, mechanical, functional control (e.g., cables, interface cards, timing)

    • topologies (star, ring, hierarchy)
    • hubs/switches and host ids
    • bridges (ARCNET, ethernet, token ring, etc.)
    • routers
    • packets (source, destination, flags, payload)
  3. Ethernet (DIX --> IEEE 802.3)

    Arbitrate access to shared communication mechanism

    1. Based on Univ of Hawaii's Aloha radio network
      • send whenever
      • await ack; no ack ==> collision
      • all detect collision, back off random amount, retransmit
    2. Bob Metcalfe did work at Xerox PARC in 1970s
    3. CSMA/CD
      • stations (multiple access or MA) listen for activity (carrier sense or CS) before transmitting
      • detect when collision occurred (CD)
      • sophisticated backoff time algorithm
    4. Popular standards
      Id Media Type Mbps Max Segment (m) Comments
      10BASE5 thick coaxial 10 500 the original
      10BASE2 thin coaxial 10 185 "Cheapernet"
      10BASE-T twisted-pair (Cat 3) 10 100 very popular
      10BASE-FL fiber optic 10 2000 most popular of original fiber
      100BASE-TX twisted-pair (Cat 5) 100 100 Fast Ethernet; auto-negotiation (10/100)
      100BASE-FX fiber (multi-mode) 100 2000  
      1000BASE-T twisted-pair (Cat 5+) 1000 100 autonegotiation of speed and duplex
      1000BASE-SX fiber (short wavelength) 1000 220  
      1000BASE-LX fiber (long wavelength) 1000 5000  
      10GBASE-T twisted-pair (Cat 6a+) 10000 100 full duplex; no CSMA/CD
      10GBASE-nnn fiber, various 10000 various  
      100GBASE-nnn   100000 various being defined now
    5. Elements of Ethernet
      1. frame -- bits of data
      2. media access control (MAC) protocol -- rules to allow shared access
      3. signaling components -- devices that send/receive signals
      4. physical medium -- cables and other hardware to carry signals
    6. Frame
      • 64 bit preamble (not often necessary)
      • 48 bit destination address
      • 48 bit source address
      • 16 bit type/length
      • 46-1500 bytes data
      • 32 bit frame check sequence/CRC
    7. MAC Protocol
      • Half-duplex mode: each ethernet-capable device or "station" is independent of others on network; connected to shared signaling channel.
      • each frame broadcast to all stations if idle channel (carrier sense; if not idle, collision detected, back off and resend later)
      • all stations check destination field for their MAC address or enabled multicast address
      • if match, NIC delivers packet to networking software
      • if no match, frame is ignored
    8. Signaling Components
      • ethernet interface
      • transceiver (may be built into interface or separate)
      • cable
      • may include repeater hub to repeat signals from one segment to another
    9. Physical Media
      • cables
      • interfaces
      • hubs -- repeaters and switches

        Repeater sends signals to all ports, and must obey round-trip timing restrictions.

        Switch remembers which port is connected to what MAC address, and sends signals only to that port.

      • bridge (connect usually dissimilar networks; e.g., ethernet and token ring)
  4. Network Protocols (data part of frame)
    • Many used in the past; most popular one now is TCP/IP.
    • Independent of physical and data link layers (e.g., Ethernet)
    • encapsulated in data part of Ethernet frame
  5. TCP/IP Basics
    1. IP and ICMP, arp and ping
      • IP: Internet Protocol

        Version 4: 4-byte addresses
        Version 6: 32-byte addresses

        Five-layer network model:

        1. Application (e.g., SMTP, FTP, HTTP)
        2. Transport (e.g., TCP, UDP)
        3. Network/Internet (e.g., ARP, ICMP, RIP)
        4. Data Link (e.g., Ethernet, Wi-Fi, PPP, EVDO)
        5. Physical (e.g., Ethernet NICs, twisted-pair cable, fiber, coax)
      • IP Datagram
        Bit Positions
        0 4 8 12 16 20 24 28
        Version Header Len. Type of Service Total Length
        Identification Flags Fragmentation Offset
        Time to Live Protocol Header Checksum
        Source IP Address
        Destination IP Address
        Options Padding
        Data starts here...

        Encapsulated in media (e.g., ethernet) frame's Data field.

      • ARP: Address Resolution Protocol

        Broadcast IP address to find MAC address associated with it. Requests are cached; use this command to see them: 

                arp -a
      • ICMP: Internet Control Message Protocol
        • uses IP datagram
        • sends message to source if datagrams sent too fast
        • detects unreachable destinations
        • redirect host to route to another gateway
        • detects if remote system is up
      • ping

        Most common ICMP datagram sender from user's perspective. Try: 

                ping 128.208.250.2
    2. UDP: User Datagram Protocol
      • Connectionless (nothing in protocol to guarantee receipt of packet)
      • Datagram
        Bit Positions
        0 16
        Source Port Destination Port
        Length Checksum
        Data starts here...

        Encapsulated in IP datagram's Data field.

      • More efficient than TCP
    3. TCP: Transmission Control Protocol
      • Connection-oriented, sequences/retries
      • Datagram
      • IP Datagram
        Bit Positions
        0 4 8 12 16 20 24 28
        Source Port Destination Port
        Sequence Number
        Acknowledgement Number
        Offset Reserved Flags Window
        Checksum Urgent Pointer
        Options Padding
        Data starts here...

        Encapsulated in IP datagram's Data field.

      • Three-way connection handshake
      • Sets up sequence numbering
      • Acknowledges receipt of segment and how much more receiver can accept
      • Maintains bytes sent and received
      • Accepts bytes of window size or tells sender to cease sending
      • Delivers data to correct application by destination port number
    4. IP addresses and notation (/x and subnet mask), routed and non-routed
      • Addresses: w.x.y.z, one byte each from 0..255
      • Classes of addresses
        First Byte Is Class Network Number Host Address
        0..127 A first byte next three bytes
        128..191 B first two bytes last two bytes
        192..223 C first three bytes last byte
        >223 D ignore ignore
      • Class A network 0 reserved for default route
      • Class A network 127 reserved for loopback address
      • In all classes, host numbers 0 and 255 reserved
        • 0 means network itself -- used by routing table to refer to entire network
        • 255 is the broadcast address -- used to deliver datagram to every host on network
      • gateway has different address for each connected network
      • subnets split network part into smaller pieces
        • allows delegation of managing network
        • can help overcome hardware differences
        • bit mask can be used to define subnet: bit on means use corresponding part of address as network 
                    IP: 192.168.10.20 Subnet mask: 255.255.255.0
          Means: Host 20 on subnet 192.168.10.0
        • alternatively, subnets can use slash notation, with number of bits that are on. The example above's subnet would be: 
                    192.168.10.0/24
      • Not all networks are routed; common "private" ones are:
        • 10.0.0./8
        • 172.16.0.0/12
        • 169.254.0.0/16
        • 192.168.0.0/16
    5. Displaying host IP address

      In a Windows CMD shell: 

            ipconfig
    6. Setting static host IP address

      In a Windows CMD shell: 

            netsh interface ip set address name="Local Area Connection" source=static addr=10.28.250.67 mask=255.255.255.0
    7. Routing or Forwarding
      • Host does its own to local and on subnet
      • Gateway(s) connected to host allow for internetwork forwarding
      • commands

        All of these will provide the same information: a display of the routing table. The last is Windows-specific. 

                netstat -nr
        route
        route print
      • Default route

        The destination is 0.0.0.0, and the netmask (or genmask) is 0.0.0.0 with the IP address of the default gateway to the rest of the world; i.e., if no explicit route is set, use the default gateway.

        Example of setting a default gateway from Windows: 

                route add 0.0.0.0 mask 0.0.0.0 128.208.250.100

        The last value is the default gateway's IP address. Or, using netsh: 

                netsh interface ip set address name="Local Area Connection" gateway=10.28.250.100 gwmetric=0
    8. Protocols and ports
      • UDP or TCP (byte 10)

        To see the recognized protocols: 

                notepad C:\WINDOWS\system32\drivers\etc\protocol
      • ports (bytes 23..24) identify applications that might be listening on port number

        To see listening ports on Windows: 

                netstat -an | findstr LISTENING
      • well-known and reserved services use ports 1..1023

        To see some of these service name to port mappings: 

                notepad C:\WINDOWS\system32\drivers\etc\services
      • ports can be dynamically allocated (source ports) 
                netstat -an
    9. Resolving domain names to IP addresses

      Construct a table of name to IP mappings ("hosts" file) or consult a domain name service (DNS).

      • hosts file 
                notepad C:\WINDOWS\system32\drivers\etc\hosts
        • Includes at least: 
                    127.0.0.1    localhost
        • Must be copied to all hosts whenever new host is added or changed
        • Backup in case DNS is down
      • Set up a DNS server address 
                netsh interface ip set dns name="Local Area Connection" source=static addr=128.95.120.1 register=NONE
        • Order of search -- use the GUI

          Can do it with wmic and nicconfig alias.

        • Must be done on all hosts whenever DNS server is added or changed
    10. Firewall

      Should set one up, but know what you want to do first.

      • ports
      • program exceptions
      • scope
    11. DHCP -- Dynamic Host Control Protocol

      Assign ("lease") a predefined IP address from a pool of addresses when some computer ("DHCP client") broadcasts a request for an address to any listening "DHCP Server". The server can also assign DNS servers and gateways.

      • when lease expires, address is returned to pool
      • when pool is empty or DHCP server cannot be found, non-routed address is used (on Windows XP's subnet 169.254.0.0/16)
      • not done on Windows XP, but on Windows Server 2003
      • common feature of Unix systems and wireless routers
    12. NAT -- Network Address Translation

      Usually used together with DHCP and a private IP address pool to map one of many private addresses to one or a few public addresses, to extend the IP addresses that can be used and protect the private addresses from some external threats.

      • not done on Windows XP, but on Windows Server 2003
      • common feature of Unix systems and wireless routers and access points
  6. Wired networks: advantages and disadvantages
    • + well-known
    • + physical security
    • + generally clear signal (no/low interference)
    • + speed
    • - cabling costs
    • - physical proximity
    • - change in speed may require re-cabling and new switches
    • - infrastructure costs (e.g., network closets)

Hands-on Time
  1. Create crossover connection (one wire, one crossover mechanism, two computers)
  2. Set ip address on both computers
    • Computer a: Use 10.208.104.m -- m is number of computer and subnet 255.255.0.0 
            netsh interface ip set address name="Local Area Connection" source=static addr=10.208.104.m mask=255.255.0.0
    • Computer b: Use 10.208.104.n -- n is number of computer and subnet 255.255.0.0 
            netsh interface ip set address name="Local Area Connection" source=static addr=10.208.104.n mask=255.255.0.0
    • Make sure Windows Firewall is off, and that the Advanced/Network Connection Settings/Settings.../ICMP/Allow incoming echo request checkbox is checked.
    • ping from a to b; ping from b to a
    • change subnet of b to 255.255.255.0
    • again, ping from a to b; ping from b to a
    • change subnet of a to 255.255.255.0
  3. check the IP configuration 
        ipconfig /all
  4. ping to www.google.com
  5. Create connection to hub (three wires, one hub, two computers, one internet connection)
  6. ping to:
    • www.google.com
    • computer a from b
    • computer b from a
  7. Set up a gateway
    • route any addresses that aren't known to the default gateway 
            route add 0.0.0.0 mask 0.0.0.0 gw 128.208.247.100
    • Check it 
            route print
    • ping www.google.com

      It won't work because you are using an unroutable IP address.

    • Change to using a dynamic IP address using the Control Panel/Network Connections.

      This also might work: 

            netsh interface ip set dns name="Local Area Connection" source=dynamic
    • ping 128.208.250.2
  8. Set up a DNS server
    • Set up the server 
            netsh interface ip set dns name="Local Area Connection" source=static addr=128.95.120.1 register=NONE
    • Check it 
            ipconfig /all
    • ping www.google.com
  9. Look at ports
    • All ports 
            netstat -an
    • All listening ports 
            netstat -an | findstr LISTENING
    • Which program is listening? 
            netstat -anob >c:\netstat.out
      notepad c:\netstat.out

      The item in the square brackets is the service name.


Cleanup
  1. Clone disk drives
  2. Remove hub and cables