The two most significant things to study on the test are the OSI model
and TCP/IP addressing and subnetting.
If you know both of those very, very well, then you are guaranteed about
50% on the exam. The rest comes from
this sheet.
The last bit of advice is that you didn’t believe me on the importance of knowing the OSI model, learn it backwards, forwards, sideways, and be prepared to answer really nit-picky questions about it.
File,
print, message, database, and applications
Determines
availability of the target host.
www,
email, ftp, telnet, edi, quake
Data
Encryption, compression, and translation services
Determines
the syntax of the data transfer.
Pict,
tiff, jpet, midi, mpeg, quicktime, etc
Dialog
control, coordinates the comunications
Nfs, sql, rpc, x windows, asp (appletalk
session protocol), DNA SCP (digital whatzit)
End-to-end
communication
Responsible for hiding the communications
from the higher layers.
TCP
/ UDP
Routing
IP /
ICMP, BootP, ARP, RARP
Routers
Framing
Ethernet II, 802.5 (token ring), 802.3, 802.2
(802.3 with dsap and ssap logical link control fields)
Media access control: MAC: 48 bits, 3 bytes
vendor + 3 bytes serial number
WAN:
High-level datalink control HDLC (cisco
default for serial links)
Synchronous Data Link Control SDLC (uses
polling)
Link Access Procedure, Balanced LAPB
x.25, slip, ppp, isdn, Frame Relay
Bridges / Switches
Wire…
v.24,
v.35, x.21, g.703, hssi, etc
Repeaters
/ Hubs
Process/application à application,
presentation, session
Host-to-host à transport
Internet à network
Network access à Datalink / physical
Ethernet
802.3
CSMA/CD | Ethernet_II | 802.2
10base2/thinnet:
185 meters
10base5/thicknet:
500 meters
10baseT: can
run above 10Mbps 200ish meters
5-4-3
rule: 5 segments, 4 repeaters, 3 segments populated
100BaseTX: 100 meters, up to 2 repeaters. Packets between 512 and 1518.
FDDI
100 Mbps
token passing
with dual counter-rotating rings
Token-ring
802.5
4 / 14 Mbps
ATM
53-byte cells
Flow Control
Contention:
CSMA/CD : Carrier Sense Multiple Access, Collision Detect
Token Passing:
Token Ring, Fiddi
Polling: SDLC,
some HDLC, some HP ethernet stuff
Store – and – forward
Buffers whole
frame before forwarding
Cut-Through
Forward frame
as soon as destination address is available
Fragment Free
Does Cut-Through after 64 bytes are received
to stop collision packets from being forwarded.
802.1d Spanning-Tree
Protocol STP
detect and
eliminates loops in routed network
STA: Spanning
Tree Algorithm
Sends
out BPDUs: Bridge protocol data units
VLAN: Virtual Lans
Create ‘logical’ networks by location, function
or department.. or protocol, or whatever.
Done with a switch using Frame-Tagging (can
be used between switches if you have ISL inter-switch link) Frame-Tagging uses unique user-defined
identifiers while within the switch fabric, and is very scalable.
Port numbers:
TCP: protocol
number 6
ftp:
21
telnet:
23
smtp:
25
UDP: protocol
number 17
Dns:
53
Tftp:
69
Snmp:
141
0-255: public
assigned
256 – 1023:
assigned to companies
1023+ user
defined / source addresses
IP Addressing:
Class leading bits decimal
range of first byte
A 0 1-127
B 10 128-191
C 110 192-223
Learn the rest of the
stuff…. Way out of scope of this.
NOTE:
Cisco considers the mask to
be the bits beyond what is ‘normal’ for that class address, not the entire
number of bits in the subnet mask, sometimes….
Protocl Stack:
Application,
presentation, session à RIP, SAP, NCP, NLSP, etc…
Transport à IPX, SPX
Network à IPX
Data link à ODL Open Data Link
Physical à whatever
IPX
Connectionless,
and communicates via sockets
Each host runs
its own internal ipx network in addition to any lan network
Addressing:
10
byte address
first
4 bytes are network – need to be unique, otherwise whatever..
last
6 bytes are node - usually just use the
MAC address
SPX
Sequence Packet
eXchange
Connection
oriented protocol
Creates virtual
circuits, with specific connection IDs
RIP
Routing
information protocol
Distance vector
protocol that establishes routes between ipx networks
Judges based on
ticks (1/18 second units) and hops
Broadcast every
60 seconds
Used to provide
each server with a complete network map
SAP
Service
advertising protocol
Servers use it
to advertise, clients use it to locate services
Broadcast every
60 seconds
NLSP
NetWare Link
Services Protocol
Link-state protcol to replace RIP and SAP someday…
NCP
NetWare Core
Protocol
Provides access
to server resources
Netware in a nutshell
-Strict
client-server model (nobody is both)
-Servers
provide files, printing, messaing, applications, and databases
-Every netware server (or cisco router)
creates a SAP table of all services offered by all servers
-A client issues a GNS (GetNearestServer
request) broadcast to find out what is on the local SAP
Netware frame types Features
802.3
default on NetWare <=3.11
802.2
default on NetWare 3.12+
ethernet_ii supports
tcp/ip and ipx
ethernet_SNAP appletalk,
tcp/ip, and ipx
--- all of these are incompatible… god love’m
RIP: Routing Information
Protocol
Distance –
vector routing protocol
Updates every
30 seconds
Route invalid
timer 90 seconds: time before route considered invalid.
Route flush
timer: 240 seconds: route removed from table
Can use metrics
(1-15) to weight against some interfaces
15 hop limit
OSPF: Open Shortest Path First
Link-state
routing
Very
infrequent broadcast updates
Extremely
granular metrics
NLSP: Novell’s Link State Protocol
Link-state
IGRP: Interior Gateway Routing Protocol
(cisco proprietary)
Distance-vector
Metrics
and hop count from 1-255
Measures
delay in units of 10 milliseconds
Measures
bandwidth – on serial connection this needs to be set, default is T1
Measures
reliability as 1-255 (255 optimal):
Measures
load: 0-255 (0 = no-load)
Allows
multi-path routing (dual links of equal bandwidth to 1 location)
Implements
hold-downs, split horizons, and poision reverse updates
Update
timer is 90 seconds, invalid timer is 270 seconds (3 times update)
Hold
down timer is 280 (3 times update +10 seconds)
flush timer is 630 seconds
(7 times update)
administrative distances:
(reliability of information)
0 = direct connection, 1 = static, 100 =
igrp, 110 =ospf, 120 = rip, 255 = unknown
EIGRP: Enhanced IGRP
Hybrid
routing protocol
Uses
distance vectors, however they are triggered by changes, not timers.
Faster
convergence, multiprotocol support
Distance- vector:
Uses
second-hand info
Problems
detecting /closing routing loops (counting to infinity)
Judges ‘best’
based on hop counts
Convergence can
get pretty lengthy
Split horizion:enforces that
inforation is not sent back in direction it came from
Route poisoning: Helps prevent incorrect
updates by setting route down explicitly
Hold-downs: prevents routes from changing too
quickly, to allow time for stabilization
Link-state:
No-second hand
info, and understands entire network
Uses LSP packets to build ‘personal’ copy of
entire network structure to route from
LSP: link-state
packets or “hello packets”
Chooses ‘best’
path based on: bandwidth, congestion, metrics, etc.
Update times
can be set very lengthy as changes cause triggered udpates.
EGP: Exterior Gateway
Protocol
Polls neighbors
exchanges info about AS with
neighbors
distance vector
very simple
BGP: Border Gateway
Protocol
Can detect
routing loops
Can work
between AS
CDP: Cisco Discovery
Protocol
Runs SNAP:
SubNetwork Access Protocol at the datalink layer.
-
this allows routers running different protocols to still communicate
-
60 second updates, 180 second hold time
#show cdp interface
interface
information, encapsulation info, and timing information
#show cdp entry <name or * for all>
shows
detailed information about other routers on network
#show cdp neighbors
shows
summary information that is being exchanged
#show cdp neighbor detail
extended
listing of info for all routers
#show cdp traffic
shows
packets sent / received amount neighbors
WAN:
POTS
Plain Old Telephone Service. (politely is: pstn: public switched
telephone network)
Demarc: end of
responsbility for provider, start for customer
CPE: customer
premise equipment
Local loop:
wire from CO to demarc
CO: Central
Office: provider POP : point of presence
SDLC
Synchronous
Data-Link Control
Used originally
for SNA
Point –to-point
or multipoint
Bounded or
unbounded media, half or full duplex
Circuit or
packet-switched networks
2 node
types. Primay or Secondary (controlled
by primary)
HDLC
High-Level
Data-Link Control protocol
This is the default communication method for
cisco routers of sync links, and is proprietary
This came after SDLC, and was modified into
LAP and LAPB.
HDLC transfer modes:
NRM: Normal Response Mode: secondaries only
speak when spoken to by pri
ARM:
Async Response Mode: sec. can speak on own
ABM:
Async Balanced Mode: nodes can be both pri and sec
LAPB:Link
Access Procedure Balanced: built into x.25
DDR
Dial on demand
Routing
·
be sure and setup route as static, and with specified distance over 200
to make sure it is ‘last resort’
X.25
Point-to-point
communication between DTE and DCE
DTE:
Data Terminal Equipment (router or terminal, or whatever)
DCE:
Data Circuit-Terminating Equipment (csu/dsi or modem)
PSE:
Packet Switching Exchange: switches inside carriers network
Addresses
defined by x.121 as a 14 digit number
DNIC:Data
Network Identified Code is first 4 bits of address
Rest
of address is assigned by the administrator
Full-duplex
protocol
‘overbuilt’
with lots of error checking
Created
SVC or PVC connections (switched or Permanent virtual circuits)
PLP
Packet
layer protocol
x.25’s
network layer protocol
modes:
call setup, data transfer, idle, call clearing, restarting
LAPB
Link
Access Procedure Balanced
Makes
sure frames are error free and sequenced
Types
of frames:
I information sequencing, flow control, error detect, recovery
S supervisory handles requests for, and suspension of transmittion
U unnumbered link setup and disconnections and error reporting
Setting
up X.25 on Cisco
#interface
<interface>
#encap
x25
#x25
address <14 digit number>
to
set the address
#x25
ips <bytes>
to
configure Input Packet Size
#x25
ops <bytes>
to
configure Output Packet Size
#x25
win <integer>
to
configure Window Input Size
#x25
wout <integer>
to
configure Window Output Size
Frame Relay
Shared
Bandwidth
Can setup a CIR
(Committed Information Rate)
Assumed
error-checking is handled at another, higher, layer
PVCs are
created at layer 2.
DLCIs: Data-link connection Identifiers : are
used to identify virtual circuit connections.
DLCI address are assigned by the provider and
then mapped to IP addresses by the router
LMI: Local
management Interface
These are
autodetected in current IOS versions….. however:
Keyword Meaning
Cisco: defined by industry group, and default
ANSI: Annex D defined by T1.617
Q933a: Defined by ITU-T Annex A Q.933A
LMI can be used
to determine the global significance of the DLCI numbers.
Setup of Frame
Relay on Cisco
#interface
<interface>
#encapsulation
frame-relay [ietf, or default is cisco]
use default to talk to other cisco routers,
use ietf encapsulation to talk to non-cisco.
#frame-relay
interface-dlci <dlci number>
to
map dlci number to current interface, or subinterface
Then
specify an IP address for that subinterface
Optionally
you can hard-code the address on the other end.
#encap
frame-relay [ietf]
#no
inverse-arp
turns
off auto addressing features
#ip
address <ip address> <subnet mask>
#frame-relay
map ip <address> <metric ? > [cisco] [broadcast]
this
lets you mix encap types, and allow broadcast over interface
Monitoring
Frame Relay
#show
frame ?
ip
ip statistics
lmi lmi stats
map map table
pvc pvc stats – this one displays the DLCI #
route route info
traffic protocol stats
PPP
Point to Point
Protocol
Userfull for
dial-up or sync links (ISDN)
Authenticate
using PAP: password authentication protocol or
CHAP:
Challenge Handshake Authentication Protocol
ISDN
Integrated
Services Digital Network
Terminal
equipment types:
TE1:
understands ISDN
TE2:
predates ISDN and needs a TA (terminal adapter) to work
ISDN reference
points:
R:
between non-isdn device and TA
S:
between terminal and NT2 device
T:
point between NT1 and NT2
U:
point between NT1 and carrier line
termination device
ISDN protocol
codes:
E:
existing telephone network
I:
concepts, terms, and services
Q:
switching and signaling
ISDN service
levels:
Connect
to lines with SPIDs (service Profile Identfiers) (phone numbers..)
BRI:
Basic Rate Interface: 2B + 1D
B
= 64kbs, D=14Kbps = 128kbs plus control
PRI:
Primary Rate Interface: 23B + 1D
Total
of 1.544Mbps
Configuring
ISDN BRI
#isdn
switch-type ?
tons
of proprietary switch types
#interface
<interface>
typicaly
bri0, or something like that.
#encap
ppp
ppp
is method used to setup isdn phone calls
#isdn
spid1 <spid1>
#isdn
spid2 <spid2>
Configuring
ISDN PRI
#controller
<interface> <location>
typicaly
something like #controller T1 1/0
#framing
efs
sets
Extended Super Framing, this is normal for T1/ PRI line
#linecode
b8zs
line-conding mechanism to assist with timing
by preventing strings of zeros
#pri-group
timeslots <value from 1-24>
sets
number of timeslots
Access lists:
Basics:
Access lists
must be created, then applied to an interface
Access lists
can filter incoming or outgoing from an interface
Packets are
compared only until a match is made
Packets that do
not meet any criteria on the list are discarded
Wildcard masking:
Nifty, acts
like a reverse subnet mask:
Example:
0.0.0.255 would wildcard all hosts on class C network
0.0.0.0
would indicate no wildcarding
keywords:
host:
0.0.0.0 : just specified host
any:
255.255.255.255: absolutely anything
* when using keyword it comes before IP
address, when using mask it comes after!!!!
List number scheme
List number range meaning
1-99
ip standard list
100-199
ip extended access list
800-899
ipx standard
900-999
ipx extended access list
1000-1099 IPX
SAP access list
IP access list creation
#access-list <number
of list> <permit | deny> <source address> [<mask>]