Pluralsight

• Contents

Terminology

From PluralSight

Physical

• 0/1

Data-Link (Frames, Switches)

• Between two physical devices

• "Wire" drawn like thunderbolt = serial

• drawn as line on logical diagram, symbol/box on physical diagram

• MAC

  • First half OUI
  • 7th MSB should be 0 (so second value cannot include values including 2)

• PDU is frame

  • dest MAC | src MAC | data | FCS
  • Dest MAC first cos quicker for destination to compare if correct
  • FCS (frame check sequence) includes CRC (cyclic redundancy check)

• Switch

  • Flood - sends frame to all devices if dest MAC unknown, but unicast
  • Broadcast - dest MAC is FF:FF:FF:FF:FF:FF, sends frame to all devices
  • Switch#show mac address-table
  • The switch is the datalink - one switch is one datalink and all devices connected to it are on that datalink (the individual connections are not datalinks)

• Broadcast and collisions

  • Broadcast domain - where the broadcast will go, switch is a broadcast domain
    • Cannot cross a router
  • Collision domain - half-duplex, frames can collide, each port on a switch/router is a collision domain, hub is a collision domain

  

• VLAN

  • layer 2 broadcast domain
  • Switch#show flash # where vlans and other settings are stored
  • Link = port

    • Access link - part of single VLAN

      sw1#sh[ow] vlan brief
      sw1#sh int[erfaces] f[astEthernet] <slot>/<port> [sw[itchport]]
      sw1#sh int trunk
      
      # create vlan
      sw1(config)#vlan <100>
      sw1(config-vlan)#name myVLAN
      sw1(config-vlan)#exit
      
      # assign port to vlan
      sw1(config)#int f <slot>/<port>
      sw1(config-if)#sw mode access
      sw1(config-if)#sw mode access vlan <100> # will create vlan if doesn't exist
      
      # delete vlan
      sw1(config)#no vlan <100> # delete
      
      # old way to create
      sw1#vlan database
      sw1(vlan)#vlan 50 name TEST
      

    • Trunk link - combine switches, carry multiple vlans

      • By default all vlans are allowed on a trunk
      • Frame tagging/trunking protocols: inter switch link (ISL), IEEE 802.1q (better)
      • Dynamic trunking protocol (DTP): dynamic desirable (cisco switch model ##50) (starts DTP frames); dynamic auto (cisco switch model ##60) (doesn't initiate DTP frames, will respond only - manually configure); ON
      • Native vlan - not tagged when sent - configure on both switches is best

      sw1(config)#int f 0/1
      sw1(config-if)#sw mode [dynamic desirable|trunk]
      sw1(config-if)#sw tr[unk] en[capsulation] <dot1q> # needed for trunk
      sw1(config-if)#sw nonegotiate # turn off DTP frames, save bandwidth if forced on
      sw1(config-if)#sw tr allowed vlan <1,100-200>
      sw1(config-if)#sw tr native vlan <150>
      sw2(config-if)#sw tr native vlan <150> # repeat commands on sw2 to mirror
      

• STP

  • Creation (for each of above)
    1. elect 1 root bridge per layered domain
       1. lowest MAC address as lowest bridge ID as lowest MAC (as priority same, default 32768)
    2. elect 1 root port per non root switch
       1. BPDUs leave root to all switches, cost increases as it enters switch
       2. once BPTUs have arrived in all ports, pick lowest cost
       3. if equal, use sender bridge ID; if equal, sender port ID
    3. elect 1 designated port per segment (link between devices)
       1. lowest root path cost
       2. note: all on root will be designated
  • Decision process
    1. lowest bridge ID
    2. lowest root path cost
       1. 10mpbs=100
       2. 100mbps=19
       3. 1000mpbs=4
    3. lowest sender bridge ID
    4. lowest sender port ID
  • bridge port data unit (BPDU fields)
    1. root bridge ID
    2. root path cost
    3. sender bridge ID
    4. sender port ID
  • bridge ID = priority (2^16, 12 bit) + MAC (48 bit)
  • stp port states
    1. blocking (20 secs)
    2. listening (15 secs) (a.k.a forward delay) (populating MAC address table)
    3. learning (15 secs) (a.k.a forward delay) (populating MAC address table)
    4. forwarding
       1. designated ports become forwarding ports
       2. non-designated (alternate) ports become blocking ports
  • per VLAN STP (PVST)
    • bridge ID = VLAN ID (4 bit) + priority + MAC
  • enhancements
    • portfast - blocking straight to forwarding (faster) (not for trunk port!)
    • BPDUguard - shutdown port if receives BPDU (from malicious switch)
  • vlan trunking protoctol (VTP)
    • vtp server mode switch
      • can modify vlan.dat
    • vtp client mode switch
      • do what server says
      • can't modify vlan.dat (i.e. make new vlan on self)
    • vtp transparent mode switch
      • not do what server says, just forward (i.e. make a new vlan (on root?), not carried over)
      • can modify own vlan.dat (i.e. make new vlan on self)
    • VTPv2 supports token ring vlan, v1 doesn't

  sw1#show spanning-tree vlan <150>
  sw1(config)#spanning-tree vlan 150 priority <n4096>
  sw1(config-if)# spanning-tree portfast
  sw1(config)#vtp mode <server|trans[parent]>
  sw1(confog)#vtp domain <domain-name>
  sw1(config)#vtp pass[word] <password>
  sw1#show vtp status
  sw1#show vtp pass[word]
  

Network (Packets, Routers)