Docstoc

Building Scalable Cisco Internetworks (BSCI)

Document Sample
Building Scalable Cisco Internetworks (BSCI) Powered By Docstoc
					           Building Scalable Cisco Internetworks (BSCI)
                                             642-901
                                           Study Guide
                                            Rob McCarty
                                              12/12/07




Contents:
Routing Protocols

-EIGRP                              Pages 2-6
-OSPF                               Pages 7-11
-IS-IS                              Pages 12-15
-BGP                                Pages 16-x

Advanced Routing Topics

-Multicast                          Pages x-x
-Route Manipulation                 Pages x-x
-IPv6                               Pages x-x
-DHCP Services                      Pages x-x




Before you read:


This guide was created with the intention to help with the Cisco exam BSCI 642-901. This is just a
quick reference, cram style study guide. It is not designed to teach the foundation topics required to
pass the exam, rather a guide to help remember specific topics. I suggest reading the Cisco Press books
associated with the exam before going through this guide, as well as getting as much hands on practice
with the Cisco IOS if you want to be successful.

This guide also assumes the reader has CCNA level knowledge to understand many of the terms used.
                                      Routing Protocols


                                                 EIGRP
Overview

-Developed by Cisco, and will only work with Cisco devices
-Considered a “Hybrid” between Link-state, and Distance vector routing protocols
-Uses backup routes called a feasible successor
-Uses the DUAL algorithm to calculate routes
-Has 3 tables: Neighbor, Routing, and Topology.
-Classless Routing, supports VLSM
-Supports multiple routed protocols ex. IPX, Appletalk, IP


Terminology

Successor: The best route, or primary route, to a network. It has the lowest FD
Feasible Successor: The backup route to a successor.
Feasible Distance (FD): The total distance to the destination network, including the AD
Advertised Distance (AD): The distance from the next-hop router, and the destination network
Active Route: Route is undergoing computation, due to a change ex. Its down.
Passive Route: Operational route
Neighbor table: List of directly connected routers
Routing table: The best routes to networks. List of successors and feasible successors
Topology table: A list of neighbors' routing tables.
DUAL: The algorithm that decides what routes to use
Reliable transport protocol (RTP): Ensures EIGRP packets reliability


EIGRP Metric

The metric is the bases for how the Successor is chosen. The lower the metric, the better the path to
that network is. It is a complex calculation that can be based on:

Bandwidth – The speed of the line
Delay – The time it takes for a packet to cross the line
Reliability – Uptime of a line
Loading – The load on the line
MTU – Maximum Transmission Unit

Although all of these can be configured, by default, only Bandwidth and Delay are used to calculate the
metric. These options are referred to as K values in the IOS.

*Note* If a metric option is configured on one router, but not another, they will NOT become
neighbors
EIGRP packets

Hello: Creates relationship with neighbors

Update: Update/send routing tables to neighbors

Query: Queries for available routes/paths to networks

Reply: Reply to a query

ACK: The RTP response to ensure the packet was received. Used for everything BUT Hello packets.



EIGRP Show Commands

show ip eigrp neighbors: Displays the neighbor table

-H (handle): Lists each neighbor in order of discovery
-Address: The layer 3 address of the neighbor ex. IP address.
-Interface: The int that the neighbor is connected through
-Hold Time: The maximum time without hearing a Hello packet before this link is considered “dead”.
The default is 15 seconds, or 3 hello packets.
-Smooth Round Trip Timer (SRTT): The average time for the router to send a hello packet, and
receive an ACK back.

show ip eigrp topology: Displays routes with their associated successors and fiesable successors (if
applicable), along with the FD to each. Also displayed here, are if any routes are in an active or passive
state.

show ip route: Displays the routing table, along with how each route has been learned.

show ip eirgp interfaces: Displays specific information for each interface that is configured for EIGRP

Peers: Number of directly connected neighbors
Pacing Time: Used to determine when eigrp packets should be sent on this interface.
Pending routes: Number of routers in the packets that are waiting to be sent

show ip eigrp traffic: Displays the number of EIGRP packets that have been sent/recv. ex. The
number of hello packets send/recv.
Load Balancing

By default, EIGRP will load balance across up to 4 equal paths to the same network. Paths are
considered equal, when they have the exact same metric.

You can however, load balance over multiple paths even if they do not have equal metrics. The
command to do this is called variance, and is run from the EIGRP configuration prompt.

ex. hostname<config-router>#variance 3

The number after the variance command is a value between 1-128, with 1 being default. This number
defines how much of a variance there can be from the metric of the successor, by a multiple of what
ever the number is.

ex. a variance of 3 means that there can be up to 3 times the metric to still be used for load balancing.
If the successor's metric is 500, then it will use any routes with up to 1500 as a metric as well.


EIGRP Default Route

A default route is a route that a router sends packets to, that it does not have a route for in its routing
table for. ex. Addresses on the Internet. In the IOS is is referred to the gateway of last resort.

Configure by setting a static route to 0.0.0.0 via 0.0.0.0 on the interface you want it to go out of.
ex. ip route 0.0.0.0 0.0.0.0 eth0/0

If you then want to advertise this route to other routers to use via EIGRP, configure the network for it
on the AS number you are using.
ex. router eigrp 1
        network 0.0.0.0

This would allow other routers use int eth0/0 on this router, as a default gateway.

You can also configure a default route on an existing route by the ip default-network command.
ex. ip default-network 192.168.0.1

This will set the current route to 192.168.0.1 as the gateway of last resort/default route.


Route Summarization

By default, EIGRP has the command auto-summary enabled. This causes the protocol to become
classfull, and does not transmit the subnet masks along with routing updates (it assumes default masks).
To turn off auto-summary so that you can use more specific and efficient summarization, simply enter
the no auto-summary command from the EIGRP configuration prompt.

You will then be able to configure specific summary routes.
ex. ip summary-address eigrp 1 192.168.0.1 255.255.255.240
EIGRP Authentication

As a way to protect the network there is the option to have EIGRP communication authenticated. It
uses MD5 hashing to encrypt a key-chain of keys which are used in the domain. Each key will have a
“lifetime” on it, which is how long the route will use that key. This way you can set 1 key to expire,
and have it automatically start using another afterwards. You can have the keys overlap, to ensure no
problems between routers when changing keys.

To configure:

key chain Key_Chain1 where Key_Chain1 is whatever name you want to call the key chain.
This takes you into the configuration of that key chain.
key 1 where 1 is the number of whatever key you want to configure.
key-string defines what the 'password' will be for this key

There are now 2 commands to run, to set how often the keys will be used.
accept-lifetime is the timeframe that this router will accept this key as valid
send-lifetime is the timeframe that this router will use this key and send to others routers.
These values are set by setting the time-to-start (hh:mm:ss, day, MONTH) and the duration, or a
specific time to stop.

Now to actually turn on authentication, you go into the interface configuration of whatever link you
want to run it. You do not need to run authentication on every link. Under the interface type:

ip authentication mode eigrp 1 md5 where 1 is the AS number

Once it is on, you need to apply the key-chain to be used

ip authentication key-chain eigrp 1 Key_Chain1 where 1 is the AS and Key_Chain1 is the name of the
key chain.

*note: If a router receives communication on a link from another EIGRP router and does not have the
correct authentication, it is immediately taken out of the routing table.


STUB Zones

A stub is a router that does not go anywhere, meaning it is not asked about routes because it has already
told its neighbors that it has nothing else. Although not unique to EIGRP, stubs help cut down on
queries within a network.

ex. A link goes down on a router that is also connected to 5 other routers. It will then query those 5
routers for a backup path, and if those routers do not have a route, will then pass the query on and on
until it finds one, or it timesout. If those 5 routers were configured as stubs, the router that lost the link
would not even ask them, as it already knows they are stubs and have no other routes.

From the EIGRP configuration prompt, use the eigrp stub command, followed by what you want to
advertise.
ex. eigrp stub static. This would advertise all/only static routes.
EIGRP Notes

-224.0.0.10 is the multicast address that EIGRP uses.
-Hello packets are NOT sent reliably, meaning there is no ACK message when they are received.
-The metric calculation is ...
-The administrative distance for Internal EIGRP is 90 and for External EIGRP 170
-A K-Value mismatch means the options being used to calculated the metric differs between connected
routers, and they will not become neighbors because of it.
-A graceful shutdown is when an interface shutdown (either administratively or due to a reload),
EIGRP will send out a final hello message with k-values of 255 to tell other routers that is it being
taken down. Also called a goodbye message.
                                               OSPF


Overview

-Link state protocol
-Uses Dijkstra's SPF algorithm
-Only sends routing updates when there are changes*
-Very resource intensive compared to other protocols

Terminology

-Neighbor Table: List of directly connected routers
-Routing Table: The current best routes to networks
-Topology Table: A complete database of all possible routes in an area
-Area Border Router (ABR): A router that connects 2 areas together, and knows about both
topologies
-Autonomous System Boundary Router (ASBR): Connects an area to a network, outside of your AS
ex A network like the Internet. Also connects to a network running a different routing protcol.
-Router ID (RID): an identifying name used by OSPF.
-Internal Router: I router that has all interfaces in one area
-Backbone Router: A router in the backbone area A.
-NBMA: Non-Broadcast-Multi-Access. ex. Frame Relay


Areas

All areas must connect back to area 0, which is considered the backbone area. Because of this, on
ABRs, one of the interfaces must ALWAYS be in area 0.

The topology table is shared and is the same between all routers in an area. ABRs, will have routes
from both areas.


Router ID

Based on the highest IP address that is active when OSPF starts on a router. However, a loopback
interface will take priority over a physical, regardless of whether it is higher or not.
ex.
int e0/0 is 192.168.1.1
int e0/1 is 192.168.1.2
192.168.1.2 becomes the RID
If you configured a loopback with the address of 1.1.1.1, then the RID would be 1.1.1.1 next time
OSPF started.

The command router-id, overrides all ip address and becomes the RID.
ex. <config>#router-id 2.2.2.2
Show Commands

show ip ospf – Displays OSPF related information, including area types
show ip ospf database – Displays details of LSAs
show ip ospf neighbor – Displays Neighbor table
show ip route – Displays routing table

When looking at the routing table, OSPF routes have 4 different tags
O – an intra-area route learned by OSPF
O IA – an inter-area router learned by OSPF
E1 – An external route.
E2 – An external route.

The difference between E1 and E2, is that E1 is the combined costs of all hops to get to that external
router, whereas E2 is just the cost the ASBR gave that route, regardless of the cost to get to that ASBR.
This is the default and should be left as unless there are multiple ASBRs to get to an external network.

Summarization

Route summarization can only be done on an ABR between areas, thus careful planning needs to be
done to ensure IP address contiguousness. Route Summarization blocks specific route LSAs from
entering the backbone and propagating to all areas. Two types of summarization can be done:

Interarea route summarization – When an ABR summarizes routes in an area
External router summarization – Summarizations of routers injected in OSPF via redistribution.
Usually done on ASBRs.

To configure interarea summarization from the OSPF configuration:
area 1 range 10.1.2.0 255.255.255.0 100, where 1 is the area ID, and 100 is the cost (optional)

To configure external summarization on an ASBR from the OSPF configuration:
summary-address 10.1.2.0 255.255.255.0 word, where word is a tag (optional) for us with route-maps.
Neighbor Communication and Packet types

Hello packets – Sent to all routers on the area. Contains neighbor table.
Database description (DBD) – Brief description of the topology, used to communicate what routes
have been learned.
Link-state request (LSR) – A request for an LSA of a route
Link-state update (LSU) – Packet that contains the LSA(s).
Link-state advertisement (LSA) -
-Type 1/Router LSA
       -A route that is flooded to all routers in an area
-Type 2/Network LSA
       -Advertises a network in an area
-Type 3/Summary LSA
       -Routers to another area. Type 1 LSAs and turned in Type 3 by ABRs
-Type 4/Summery LSA
       -Like type 3, but involve ASBRs
-Type 5/External LSA
       -An advertisement of another AS by an ASBR
-Type 7
       -Use in NSSA to pass LSAs from external networks

Link-state acknowledgment (LSACK) – Acknowledgment message that ensures reliability of updates
Neighbor process


OSPF Metric/Costs

OSPF makes its routing decisions based on the cost of routes. The lower the cost, the better, or faster,
the route.

Cost = 100 / Bandwidth in Mbps

ex.
FE 100Mbs = 1
Ethernet 10Mbps = 10
T1 1.544 = 65
56k = 1785

However, by default, because there cannot be a fraction of a cost, anything over 100mbps, would also
get a cost of 1, no matter the speed.

There are two ways to change the default costs.

First, you can just manually specify what you want the cost to be. To do this, use the command
ip ospf cost on an interface

Or, you can change how the cost is calculated so that it adjusts to the bandwidth more accurately.
Under the OSPF configuration, use auto-cost reference-bandwidth 100 , where 100 is the metric you
want to use in Mbs (100 being default)
DRs and BDRs

Anytime there is a shared segment, there will be a DR and a BDR. P2P connections however, do not
elect or need this. All other routers on the segment, will be tagged as DR-Other

All updates that are needed to be sent out on that segment, will be sent out by the DR.

The DR and BDR are elected based on the router priority that is sent in the hello packet. Since by
default all the router priorities are set to 1, the tie break is the higher RID.

You can change the router priority, to ensure a router becomes a DR or BDR, use the command router-
priority
ex. <config>#router-priority 2. If all other routers were left as 1, this would become the DR. You can
set the priority to 0 if you would like to disable the router from ever becoming DR/BDR

Authentication

Two types of authentication are supported in OSPF. Simple password (plain text), and MD5
authentication. Passwords are inserted into the header of OSPF packets, and neighbors must share the
same passwords on the interface in which they are connected. To configure authentication from the
interface config:

ip ospf authentication-key word, where word is the password you want to use, or for an MD5 key, use
ip ospf message-digest-key 100 md5 word, where 100 is the key-id, and word is the password.
ip ospf authentication for simple password, or ip ospf authentication message-digest for MD5
encryption of the password.


Default Route

OSPF routers do not, by default, generate a default route. To configure:
default-information originate always. If there is already a default-route in the routing table, this will
tell this router to advertise it to others. If there is not already a default-route, the always (optional)
option at the end forces it to advertise the 0.0.0.0 anyway.

NBMA Types and Configurations

Broadcast, multi-access networks:
Ex. Ethernet.
Single operation mode. Uses DR/BDR and 10 second hello intervals

Point-to-Point networks:
Ex. ISDN
Single operation mode. No DR/BDR and 10 second hello intervals.

Non-Broadcast, Multi-Access (NBMA) networks:
Ex. Frame Relay
Five operation modes.
Virtual Links and Configurations

Virtual links are used to connect areas to the backbone, through other areas, without actually sharing an
ABR with area 0. You can also use it to connect discontinuous area 0s in an AS. However, this should
only be considered a temporary solution, and not part of the design. To configure:

area 1 virtual-link 2.2.2.2, where 2.2.2.2 is the router ID of the remote router you want to neighbor
with.


Stub, Totally Stubby Areas and Not-So-Stubby-Areas

Stub area – The ABR does not forward external routes into this area. In routers need to reach external
networks, they use default routes.

Totally Stubby area – This area does not accept external routes, or summary routers from other areas.

NSSA – A stub area that allows routes to still come in under LSA type 7. This areas can also have
ASBRs.

Stub and Totally stub areas should have only 1 exit, or only 1 default route to exit. All routers within
an area must be configured as stub routers before they will become neighbors.

To configure stub areas, #######


Notes

*updates are still sent every 30mins to neighbors, to ensure convergence.
LSAs on virtual links do not age, and will not be refreshed every 30mins
                                                  IS-IS


Overview

-A router is referred to as an intermediate system, or an IS.
-Is a Link-State Routing protocol
-Uses SPF algorithm
-Originally created to be used with the OSI protocol
-Integrated IS-IS was created to support TCP/IP
-Integrated IS-IS still requires and OSI address, which is now called a CLNS, which acts as the layer 3
protocol for IS-IS.


Terminology

Connectionless Network Services (CLNS): The OSI address used by IS-IS
L1 Router: Only knows of networks in their own area
L2 Router: Only knows of routes on the backbone.
Network Service Access Points (NSAP): The CLNS address that is applied to a router.
NSAP Selector (NSEL): When this is set to 0, the SNAP is called a NET
Network Entity Title (NET): Used to identify an OSI host in a domain
PDU
LSP


IS-IS Routing

There are 3 types of routers. L1, L1/L2, and L2. Areas connect to each other via links instead of
ABRs. However, L2 routers must be contiguous throughout the network, and have a physical path
through each area.

Because IS-IS runs through CLNS, it is less resource intensive than OSPF. Fewer updates are sent, and
there are much fewer update types to be sent.

Metrics

The default on all connections, is a metric of 10, regardless of the connection type or speed. To make
IS-IS more efficient, you can configure the metric on a link:

isis metric 30 level-2, where 30 is the new metric value, and level-2 is the interface types to apply to.

The metric can be set to anything from 1 to 63
CLNS addresses

Unlike IP, CLNS can apply to an entire router, as opposed to a specific interface. IS-IS requires CLNS
addressing on all nodes for them to discover each other.

CLNS addresses that are used by routers are call NSAPs. NSAP addresses contain the OSI address,
and a link to the higher-layer process. It is the equivalent of the IP address and TCP.

The NSAP address is broken down into 3 parts; the area address, the system ID (6 bytes), and the
NSEL (1 byte), with a maximum of 20 bytes total.

The first part of an NSAP address is AFI. If this is set to 49, it means that the address is for internal use
(private). An example of an address:

49.0004.0000.0c4b.7b3c.00, where 49 is to specify its private, 0004 is the area ID, system ID is the
MAC address, and the NSEL set to 0, making it a NET address..

The area address is associated with the IS-IS routing process. All routers in an area must be set the
same, and is what defines an area, and is what is used in L2 routing. L1 routers only see routers with
the same area address.

The system ID must be unique within an area. Usually use the MAC address, but anything 6-bytes
long is possible. L1 routing is based on this ID, and must be unique.


Route leaking

Because L2 and L1 computations are separate, it is possible to have asymmetric routing. Meaning the
path a packet takes to reach a destination, may take a different path on the way back, due to L2 and L1
routers seeing different paths to take.

Route leaking is like redistribution, where the L2 routers share some interarea routes with L1 routers.


IS-IS packets

OSI recognizes a frame as a data-link PDU, and a packet as a network PDU. The 4 other types of
PDUs are:

Hello PDU (ESH, ISH, IIH) – use to establish and maintain adjacencies.
LSP – Used to distribute link-state information
PSNP – Used to ACK or request missing link-state information
CSNP – Used to describe an entire LSDB

L1 and L2 information is distributed in separate LSPs. LSPs in a broadcast network are sent out as
multicast, and as unicast on P-t-P networks.

The default interval on Hello packets (IIH) is 10 seconds. The dead timer is still 3 times the hello
timers
Network types

IS-IS only works in with broadcast networks, or Point to Point. It does not know about NBMA
networks, and it is recommended to use point-to-point over NBMA clouds. IS-IS assumes that LANs
and WANs are in a full mesh state, and broadcasting can be achieved.


LSDB synchronization

When an LSP is received, it is then flooded to all neighbors except the one it came from. L1 LSPs in
their local areas, and L2 LSPs through the backbone. A sequence number is used to ensure only current
updates are listened to.

Adjacent routers exchange CSNPs to compare their LSDB, and are used to determine whether their
LSDBs are synchronized. PSNPs are used to acknowledge receipt of LSPs and to request anything
missing.


Configuring IS-IS

To enable IS-IS on a router:
router isis 10, where 10 is the area-tag. The area-tag is not like an area ID from OSPF, rather it is more
like the process-id. If it is omitted, the IOS assumes 0 in its place.

To set a NET for a router once IS-IS has been enabled:
net 49.0004.0000.0a2c.93bc.00

To set the router type:
is-type level-1 or level-1-2 or level-2-only. It not set, the default is a L1/L2 router.

Once you have set these two things, you then need to go into the interface you want IP routing:
ip router isis 10, where 10 is the area-tag.

To change the router type on a per-interface basis:
isis circuit-type level-1 or level-1-2 or level-2-only.


Show commands

show clns – displays general info about CLNS network
show clns 10 protocol – specific info on an area, where 10 is the area tag.
show isis 10 route - displays L1 routing table, where 10 is the area tag.
show clns route – displays L2 routing table
show isis 10 database – displays the LSDB
show isis 10 topology – displays L1 and L2 topology tables
Summarization

To configure route summarization from the router config:
summary-address 192.168.1.2 255.255.255.0 level-1-2
                                                 BGP


Overview

A routing protocol designed for the internet to connected AS.
An InterDomain Routing Protocol (IDRP).
The routing table on a core internet BGP router can be more than 200,000 routes.
Considered a 'path vector' type of protocol
Because BGP is so scalable, it has very slow convergence time.
Uses TCP for reliability

Terminology

Interior Gateway Protocol (IGP):
External Gateway Protocol (EGP):
Autonomous System (AS):
External BGP: When BGP is running between routers in different AS
Internal BGP: When BGP is running between routers in the same AS
Peer Group: A group of neighbors configured to have the same update policies.


Multihoming

Multihoming is when an AS has more than one connection to the Internet. This could be for reliability,
or for performance reasons. BGP helps in preventing routing loops, as well as being able to configure
which links to use, and when. There are 3 ways to accomplish this

-Each ISP passes only a default route to the internal routers.
Path manipulation cannot be done by the AS, as the ISP controls what default route is used.

-Eash ISP passes only a default route and ISP specific routes
This allows more optimal routing decisions to be made, and allow routers to router through better paths
(if they exist) to specific networks.

-Each ISP passes all routes to the AS
Although resource intensive, this all allows the best routes to be chosen to any network.


BGP synchronization

BGP synchronization is disabled by default. With synchronization on, BGP routers in a transit path do
not use or advertise routers to external AS, thus it is unnecessary to turn it on. Redistribution from
BGP into an IGP is not scalable due to the size of some BGP tables, thus most networks will run full-
mesh IBGP networks.
Packet types

Once TCP is established between 2 peers, each side sends an open message. A keepalive is sent in
reply if the open is acceptable. Once this has happened, updates and notifications can be exchanged.

Open message - Includes the version of BGP running, AS number, hold time, BGP router ID (which is
chosen the same way as OSPF), and any optional options.

Keepalives – A simple message header to keep the hold timer from expiring.

Updates – Updates only one path at a time. May also include withdrawn routes, path attributes, and a
list of IP address prefixes that can be reached by this route.

Notifications – Sent when BGP routers detect an error on a route.


Neighbors

Also known as peers. BGP routers do not form relationships with all routers, rather a select few in
which it exchanges updates with. Peers can be either internal, or external of the AS. Peers are
configured manual, and do no discover each other automatically.

There are 6 states for a neighbor to be in:

Idle
Connect
Active
Open Sent
Open Confirm
Established

Only when routers are in an established state, will updates, keepalives and notification messages
exchanged.


Attributes

Attributes are what the routing decisions are based on, like a metric. An attribute can be:

well-known mandatory
-AS-path, type code 2.
A list of AS numbers that a route has traversed through.

-Next hop, type code 3
The next hop IP address to get to a destination.

-Origin, type code 1
Describes where the route came from. Can be one of 3 options: IGP, EGP, or incomplete(unknown)
well-known discretionary
-Local preference, type code 5
Indicates which path is preferred to exit an AS.

-Atomic aggregate
type code 6

optional transitive (partial)
-Aggregator
type code 7

-Community, type code 8
A way of tagging routes for filtering/maps

optional non-transitive
-Multiexit-discriminator (MED), type code 4
Also called the metric, indicated to external neighbors the preferred path into the AS

Well-known – all BGP implementations must recognizes and propagate. Mandatory must appear in
updates, and discretionary does not.

Optional – BGP routers might propagate the attributes. Transitive routers will pass the attributes on as
partial, and non-transitive will delete the attribute.

Weight – A cisco specific attribute that applies only to the local router. Routes with a higher weight are
preferred.


Route Selection

Multiple paths may exist to a network, and they are all kept in the BGP table. However, BGP select
only one path to get to a network, which goes into the routing table. The decision is based on the
attributes set. BGP does not consider routers with AS loops, or unreachable next-hops.

Configuration of BGP

To enable BGP on a router, use
router bgp 65100, where 65100 is the AS. You can only have 1 AS on a router.

To activate a session with a BGP neighbor use:
neighbor 192.168.0.2 remote-as 65101, where the IP is the address of the neighbor

To create a peer group, use the command
neighbor group1 peer-group, where group1 is the name of the group.
neighbor 192.168.0.2 peer-group group1, with the IP address of the neighbor to add to a group

To use a loopback address when configuring a neighbor, use update-source option:
neighbor 1.1.1.1 update-source loopback 1, where 1 is the interface-number
                                            IP Multicast




Overview

Networks can be more efficient by sending the same information once, to multiple nodes
Use Class D addresses
Most applications will use UDP


Terminology

IGMP
CGMP
PIM
RP
RPF


Multicast IP and MAC addresses

Multicast addresses are in the class D range, 224.0.0.0 to 239.255.255.255. Some common local scope
IP addresses reserved by IANA

224.0.0.1 – All hosts
224.0.0.2 – Al multicast routers
224.0.0.5 – All OSPF routers
224.0.0.6 – OSPF DRs.
224.0.0.9 – RIPv2 routers
224.0.0.10 – All EIGRP routers

In layer 2 environments, the IANA has reserved a block of MAC addresses starting with 01:00:5E to be
marked as multicasts.


IGMP

Used between hosts and their local router to join multicast groups, so that the router knows to forward
multicasts to the them. IGMP snooping is a way to prevent multicast flooding on layer 2 switches,
since they cannot see the IP address of hosts that have joined the multicast.
The current version is IGMPv3. It allows for hosts to request access to multicast sessions, as well as
send out leave-group messages to stop the router from sending the session. V3 also adds the ability to
filter multicasts based on the source so that hosts can have control over what is received.

By default, layer 2 switches flood IGMP groups. One solution is to manually add multicast mac
addresses, but this is not scalable.

IGMP snooping is a way for switches to listen to all IGMP packets, and update their tables respectively,
essentially becoming Layer 3 aware. This can be resource intensive.


CGMP

Cisco proprietary protocol that runs between a router and a switch. When a host joins IGMP group,
the router creates a CGMP packet with the mac address of the client that requested for the switch, so
that the switch can forward the traffic appropriately. The multicast mac address that the switches
listens on is 0x0100.0cdd.dddd.


PIM routing

PIM is independent of the routing protocol and uses its own routing table, but gets its information for it
via the unicast routing protocol running. Multi-cast routers use PIM to dynamically creates distribution
trees to determine the path of multicast traffic. There are 2 types of these trees.

Source Tree – Direct paths to each multcicast group from the source.

Shared Tree – A single tree shared by all routers. I has a single common root called an RP which all
sources forward to.

Multicast routing also uses Reverse Path Forwarding, which does not send a packet to a receiver, but
instead sends the packet away from the source.

There are 2 modes that PIM can be in:

Sparse mode (PIM-SM) – Uses a 'pull' model to send traffic. Uses the shared tree. Sources register
with the RP

Dense mode (PIM-DM) – Uses a 'push' model to flood traffic. Uses the source tree. Routers that do
not want the multicasts, send requests to the source to prune them from the tree.

There can also be the option of PIM sparse-dense mode that is a hybrid of the two, that allows for both
groups concurrently. Cisco recommends this option to always be used.
Configuration

To enable multicast routing on a router, use
ip multicast-routing command

To enable a mode, from an interface:
ip pim sparse-mode or ip pim-sparse-dense mode.

To set the router to be the RP:
ip pim send-rp-announce ethernet 0/1 scope 3 group-list 101, where ethernet 0/1 is the interface to run
on, 3 is the TTL maximum hops to announce to, and 101 is the access-list of hosts to send to. The
announcements are sent on the 224.0.1.39 address.

To manually listen on interface to a multicast group:
ip igmp join-group 224.0.0.55, with the IP of the group to listen on.

To join an interface to a group so that it automatically forwards those packets:
ip igmp static-group 224.0.0.55


Show commands

show ip mroute – displays the IP multicast routing table

show ip pim interface – displays information about routers configured for PIM

show ip pim neighbor – displays discovered PIM neighbors

show ip pim rp – displays RP information

show ip rpf – displays RPF information for the RP

show ip igmp interface – displays multicast information for that interface

show ip igmp groups – displays the multicast groups this router is part of
                                                 IPv6



Overview

-A new form of Internet Protocol, designed to replace IPv4 (current)
-No broadcasting/boardcast address
-IPsec standard on all communication
-around 300,000,000,000,000,000,000,000,000,000,000,000,000 available addresses


Formatting

IPv6 addresses are 128bits long. The are broken up into 8 groups, of 4 hex characters each.

ex. 2001:0002:0349:0000:0000:0000:0A56B:98AA

To make them shorter, and more manageable, two things are allowed.

First, any groups of consecutive zeros, can be dropped once, and replaced with ::

ex: 2001:0002:0349::0A56B:98AA

Second, any leading zeros on any groups, can be omitted

ex: 2001:2:349::A56B:98AA


Types of Communications

Unicast: Same as IPv4, sent to one host.
Multicast: Sent to many hosts listening to a multicast address, and replaces broadcasts.
Anycast: Sent to the closest host

All devices will usually have more than one IP address. The types of addresses are :

Link-Local: Layer 2 domain. ex. Like a local segment

This address is assigned automatically without DHCP, a lot like the 169.254.x.x address in IPv4. It is
used to locally communicate with other devices on the segment, without needing a DHCP or assigned
addresses. Every device will always at least have this.
All link-local address will start with FE80. The first 10 bits, will always be 1111 1110 10, followed by
54 bits of zeros. This way you can look at an address, and know right away that it is link local

The last 64 bits of the address, will be the MAC address of the device, with FFFE in the middle.

ex. If the mac address was 0012:21D5:012E, the link-local address will be

FE80:0000:0000:0000:0012:21FF:FED5:012E or FE80::12:21FF:FED5:12E short-formed.


Unique: Organization wide addressing

Used within an organization to define network boundaries, or domains. They can be compared to the
private IP addressing of IPv4 (10.x.x.x, 172.16.x.x etc)

The first 8 bits in a Unique address, are 1111 1101, which equals FD.

The next 40 bits represent your global ID, which is your organization's ID. This is unique per
organization, and is assigned to all devices within.

The next 16 bits will be the subnet ID

The last 64 bits will be the interface ID. This can be manually assigned, or automatically by a DHCP

Global: Internet address

This will be the address that lets you route on the Internet. You can still use a NAT with a unique
address internally, but Internet routers will route by the Global address.

The first 3 bits must be set to 001, making all addresses start with 2. The next 45 bits or less, will
comprise the Global Routing Prefix. The internet is expected to use addresses starting with 2001::

The next 64 bits (less what ever the global used), will be the subnet ID, to narrow down the location of
the route.

The last 64 bits, will be the interface address.




Multicast Addresses

First 8 bits are set to 1111 1111, which means they start with FF

The next 4 bits, are called the FLAG. This is used to set some basic options, like whether the address
is permanent (RFC defined) or temporary (something you create).

The next 4 bits define the scope. This specifies how far this multicast goes. The options are
1 – Interface (loopback)
2 – Link
3 - Subnet
4 - Admin
5 - Site
8 - Organization
E – Global (internet)

The last 112 bits are the address. This can be set to whatever you want, as long as they are not already
assigned to something else.

Some of the already assigned addresses are as follows:

FF02::1 - All nodes on the link (pretty much a broadcast)
FF02::2 - All routers on the link
FF02::9 - All RIP routers on the link
FF05::101 - All NTP servers in the site
FF02::1:FFxx:xxxx - “arp” msg to all nodes on the link, where the last 6 digits match the x's.
ex. FF02::1:FF00:28FD would be a way to find the layer 2 address of a system that ends in 00:28FD.
                                           DHCP Services

Overview

Used to provide clients with IP addresses dynamically


Terminology

bootp
DHCP binding


Configuration

DHCP service and relay agent are enabled by default. To turn them on use
service dhcp command

To specify a database to store DHCP bindings:
ip dhcp database \\server1

To disable conflict logging
no ip dhcp conflict logging

To specify an exclude list
ip dhcp exclude-address 192.168.0.1 192.168.0.10

To create the DHCP pool to be used and enter specific DHCP config:
ip dhcp pool mypool
To specify the range to be used
network 192.168.1.0 255.255.255.0
dns-server 192.168.0.2 /24 to specify DNS
default-router 192.168.0.1 /24 to specify a gateway


Relay agent

A relay agent relays DHCP requests across domains. Because DHCP requests are broadcasts, this
needs to be on in order to reach a DHCP server through a router.

In order for a relay agent to work, the ip helper-address needs to be assigned to an interface. This
changes the broadcasts into a unicasts and forwards them to the correct server.
Show commands

show ip dhcp database – displays recent activity in the database
show ip dhcp binding – displays a list of bindings
show ip dhcp conflict – displays any address conflicts
clear ip dhcp binding – clears bindings

				
DOCUMENT INFO
Description: Channel Partner, Cisco Certified Network Professional, Course Objectives, troubleshooting network, how to, enterprise network, Interconnecting Cisco Network Devices, Cisco Career Certifications, Cisco Routing, Microsoft Word, Study Guide, Wireless Networking, Wireless LAN, Cisco Certified Network Associate, Todd Lammle, CCNA Cisco Certified Network, ccna exam, Cisco CCNA, exam objective, wireless clients, Practice Exam, Wireless Training, Certification Exam, Cisco certification, Exam Questions,