These notes are for myself.  Some are sort of cryptic.  They may or may not help you out.  They are for the Cisco Certified Network Associate exam.

OSI networking model: 1. Physical layer (bits)     Electrical, mechanical, procedural, and functional specs, i.e. Ethernet, Token Ring, etc. 2. Data Link layer (frames)     MAC (802.3), LLC (802.2), etc. 3. Network layer (packets or datagrams)     IP, etc. 4. Transport layer (segments)     TCP, UDP, etc. 5. Session layer (data)     NFS, SQL, RPC, X11, Appletalk session protocol, DNA SCP, etc. 6. Presentation layer     PICT, TIFF, JPEG, MPEG, etc. 7. Application layer     www, ftp, email, etc.

DDN/DOD networking model: 1. Network access layer     Equates to OSI Physical and Data Link layers 2. Internet layer     Equates to OSI Network layer 3. Host-to-Host layer     Equates to OSI Transport layer 4. Process/Application layer     Equates to OSI Session, Presentation, and Application layers

Cisco's 3 layer networking model: 1. Core layer     Core of the network, high data rates, FDDI, ATM, Sonet, Fast Ethernet 2. Distribution layer     Workgroup layer, routing, filtering, separate broadcast domains, and WAN access 3. Access layer     Desktop layer, separate collision domains, hubs, switches

Subnetting:

Binary bits (big endian): 128 64 32 16 8 4 2 1
                 example:   1  0  0  0 0 0 0 1 = 129

Bits                         Subnet           Hosts *      Wildcard (reversed!)
00000000 =   0       0    255     00000000 =   0
10000000 = 128       0    128     00000001 =   1
11000000 = 192       2     64     00000011 =   3
11100000 = 224       6     32     00000111 =   7
11110000 = 240      14     16     00001111 =  15
11111000 = 248      30      8     00011111 =  31
11111100 = 252      62      4     00111111 =  63
11111110 = 254     126      2     01111111 = 127
11111111 = 255     254      0     11111111 = 255

*  Routers can enable subnet zero to increase the host count by 1 (WARNING,
old BSD based UNIX systems use subnet 0 for broadcast!  If you have old BSD
systems on your network, do not enable subnet 0)

CIDR addressing (RFC 1519, Introduction date: 1993)
The new (generally accepted around 1997 I guess) way of showing network
address ranges is CIDR addressing (Classless Inter-Domain Routing) which
is written as a slash and a number, i.e. "/24".  The slash number represents
how many bits are in the netmask.  Thus, a /24 is 24 bits which is a
"classic" Class C address space (the bits go from left to right [big endian]).
CIDR: 11111111.11111111.11111111.00000000
mask:    255  .   255  .   255  .   0
range: x.x.x.0 - x.x.x.255  (remember, everything is octal so 2^8 - 1 is 255)
example: 10.0.0.0 - 10.0.0.255 (A "classic" Class C address space)

There are alot of weird address allocations now that CIDR is being
used (the old Class networks are no longer used for the most part).
For instance, a company may have a /23 address space which is:
CIDR: 11111111.11111111.11111110.00000000
mask:    255  .   255  .   254  .   0
range: x.x.x.0 - x.x.1.255
example: 10.0.0.0 - 10.0.1.255 (assuming the first subnet of /23)

A Class A network (/8, 255.0.0.0) can have 2^24 - 2 = 16,777,214 hosts
A Class B network (/16, 255.255.0.0) can have 2^16 - 2 = 65,534 hosts
A Class C network (/24, 255.255.255.0) can have 2^8 - 2 = 253 hosts

It's amazing how much address space is wasted.  I once consulted at a
company that had a full Class A allocation, but only had around 80,000
computers in the entire company.  (They were also really stupid and put
live addresses on each desktop instead of using private address space
and NAT).
 

Calculations: 256 - netmask = subnetwork
example: 256 - 240 = 16 so subnetworks are: 0,     16,      32,      48, etc.
                             broadcast are:        15,      31,      47, etc.
                           host ranges are:      1-14,   17-30,   33-46, etc.

Number of subnets: 2^x - 2 where x = number of mask bits
example: mask of 192 is 2 bits, 2^2 - 2 = 2 subnets

Number of hosts: 2^x - 2 where x = number of unmasked bits
example: mask of 192 is 6 unmasked bits 2^6 - 2 = 62

Administrative distance:

0         Directly connected interface
0         Static route pointing to an interface
1         Static route pointing to a next-hop neighbor
5         Summarized EIGRP route
20       External BGP route
90       Internel EIGRP route
100     IGRP route
110     OSPF route
115     IS-IS route
120     RIP route
140     EGP route
170     External EIGRP route
200     Internal BGP route
255     Unknown routing source

Two types of links in a switched environment: Access and Trunk links.
-  Access links:  Links that are only part of one VLAN.  VLAN information is removed from frame before it goes out the port.
-  Trunk links: Trunks carry multiple VLANs.  Used to connect to other switches, routers, or servers.
     - Two types of Trunk framing: ISL (Cisco only) and 802.1.q
     - Can carry 1 to 1005 VLANs
- Frame tagging:  assignes user-defined ID to each frame, sometimes called a VLAN ID.
-  Frame types:  Inter-Switch Link (ISL) Fast Ethernet and GigaEthernet only.  Cisco Only
                             - Adds a 26 byte header and a 4 byte FCS, can create a GIANT packet (1522 bytes)
                             IEEE 802.1.q  IEEE standard
                             LAN emulation (LANE) used with ATM
                             802.10  used with FDDI
- VLAN Trunk Protocol (VTP)
     Server, client, and transparent server
     Advertisements are sent every 5 minutes or whenever there is a change
     VTP pruning: Updates are onyl sent to devices on the VLAN that changed.  Off by default.

r2509#sh cdp nei
Capability Codes: R - Router, T - Trans Bridge, B - Source Route Bridge
                  S - Switch, H - Host, I - IGMP, r - Repeater

Device ID        Local Intrfce     Holdtme    Capability  Platform  Port ID
r2514               Eth 0          162           R        2500      Eth 0
r2504               Ser 1          175           R        2500      Ser 0
r2501               Eth 0          178           R        2500      Eth 0

r2509#sh cdp nei detail
-------------------------
Device ID: r2514
Entry address(es):
  IP address: 10.0.0.14
Platform: cisco 2500,  Capabilities: Router
Interface: Ethernet0,  Port ID (outgoing port): Ethernet0
Holdtime : 124 sec

Version :
Cisco Internetwork Operating System Software
IOS (tm) 2500 Software (C2500-JS-L), Version 11.2(17), RELEASE SOFTWARE (fc1)
Copyright (c) 1986-1999 by cisco Systems, Inc.
Compiled Mon 04-Jan-99 17:27 by ashah

-------------------------
Device ID: r2504
Entry address(es):
  IP address: 172.16.1.2
Platform: cisco 2500,  Capabilities: Router
Interface: Serial1,  Port ID (outgoing port): Serial0

--<cut>--

Ethernet_802.3    up to NW 3.11
Ethernet_802.2    since NW 3.12
Ethernet_II       supports both TCP/IP and IPX
Ethernet_SNAP     Appletalk, IPX, and TCP/IP

Interface       Novell frame type     Cisco Keyword
---------       -----------------     --------------
Ethernet        Ethernet_802.3        novell-ether (default)
                Ethernet_802.2        sap
                Ethernet_II           arpa
                Ethernet_snap         snap
Token Ring      Token-Ring            sap (default)
                Token-Ring_snap       snap
FDDI            Fddi_snap             snap (default)
                Fddi_802.2            sap
                Fddi_raw              novell__fddi

IPX routing is easy:

r2509#conf t
r2509(config)#ipx routing
r2509(config)#int e0
r2509(config-if)#ipx network 10
r2509(config-if)#int s1
r2509(config-if)#ipx network 10a
r2509(config-if)#end
r2509#sh ipx route
Codes: C - Connected primary network,    c - Connected secondary network
       S - Static, F - Floating static, L - Local (internal), W - IPXWAN
       R - RIP, E - EIGRP, N - NLSP, X - External, A - Aggregate
       s - seconds, u - uses

2 Total IPX routes. Up to 1 parallel paths and 16 hops allowed.

No default route known.

C         10 (NOVELL-ETHER),  Et0
C        10A (HDLC),          Se1

Note: "the 1 parallel paths".  To enable load balancing, enter
    r2509(config)#ipx maximum-paths 2
  To enable load balancing on a per host basis, enter:
    r2509(config)#ipx per-host-load-share

IPX subcommands:
r2509#sh ipx ?
  access-lists  IPX access lists
  accounting    The active IPX accounting database
  cache         IPX fast-switching cache
  compression   IPX compression information
  eigrp         IPX EIGRP show commands
  interface     IPX interface status and configuration
  nasi          Netware Asynchronous Services Interface status
  nhrp          NHRP information
  nlsp          Show NLSP information
  route         IPX routing table
  servers       SAP servers
  spx-protocol  Sequenced Packet Exchange protocol status
  spx-spoof     SPX Spoofing table
  traffic       IPX protocol statistics

IP extended can be based on IP proto, eigrp, gre, icmp, igmp, igrp, ip, ipinip, nos, ospf, tcp, udp, ahp, esp

LMI   Local Mangement Interface.  Types: cisco, ansi, q933a

ISDN

NOTE: A U interface has a built in NT1, A S/T interface does not.  (some Sun computers have a built in S/T.  Also, SGI Indy's have a S/T.)

  North America uses U interface (2 wire).  Needs NT1 to convert to 4 wire S/T
  TE1 Terminal Equipment type 1.  Can plug right into ISDN network
  TE2 Terminal Equipment type 2.  Older equipment, needs TA
  NT1 Network Termination 1 implements the ISDN physical layer for user
  NT2 Network Termination 2 is the provider's equipment, PBX, etc
  TA  Terminal Adapter converts TE2 to TE1 wiring.

  R reference point. Defines point between non-ISDN equipment (TE2) and TA.
  S reference point. Defines point between customer's router and a NT2
  T reference point. Defines point between NT1 and NT2.  S and T can be the same.
  U reference point. Defines point between NT1 and line-termination equipment (NT2) (No TA/NT1 needed).

  ITU Protocols: protocols beginning with:
  E  deal with ISDN on existing network (POTS)
  I  deal with concepts, aspects, services
  Q  deal with switching and signaling

  isdn q921  layer 2
  isdn q931  layer 3