Вы находитесь на странице: 1из 10

EIGRP Fundamentals:

 Used to be Cisco Proprietary IGP. Now, Cisco opened it up to 3rd party vendors as
well.
 Very Fast Convergence. Let us say we have more than one path to the destination
and if the primary path fails, it could converge quick and use the backup path.
 Very scalable
 Load balancing over unequal cost links. Using variance option. OSPF cannot do.
 Classless VLSM support. We can advertise information about the subnet.
 Multicast destination IP: 224.0.0.10
 Administrative Distance – 90.

Components of EIGRP:

 Neighbor discovery. EIGRP speaking routers can automatically discover other


EIGRP speaking routers.
 Reliable Transport Protocol (RTP). Ability to guarantee that the EIGRP packets
were sent to the neighbor and in order. Provide reliability when needed.
 Diffusing Update Algorithm (DUAL). To determine the most attractive route
(successor) to a destination network that is known to an EIGRP speaking router. An
EIGRP neighbor that can get us to the destination network without causing a routing
loop and acts as a backup to a successor router (Feasible Successor).
 Protocol Dependent Modules – Developed with multi-protocol support in mind.

EIGRP Neighbourship Parameters:

 Corresponding interfaces must be able to send and receive packets.


 IP address in same subnet at either ends.
 AS number must be the same.
 K values must match.
 Connected interfaces must not be defined as PASSIVE.
 Authentication must pass.

EIGRP Data Structures:

 Neighbor table – Contains information about EIGRP speaking neighbors.


 Interface table – Contains information about EIGRP speaking interfaces.
 Topology table – List of Routes known to EIGRP.

(2172416 / 28160) – Feasible Distance / Advertised Distance. Metric to reach the


destination – Advertised Distance. When you add up to the cost to that distance – grand
total – Feasible Distance.

EIGRP Timers:

 When interface goes down, router realizes the networks advertised earlier by that
corresponding other end router is not anymore reachable through that device and
the action is immediate.

When interface is up/up but the how to determine if the EIGRP speaking neighbor is still up
or not?
 HELLO Timer – 5 seconds (LAN) – Tells how often should I send HELLOs
 HOLD Timer - 15 seconds – Tells my neighbor how long it needs to wait for
receiving a HELLO.

Timers are adjustable. Timers need not match with the neighbors unlike OSPF. HOLD Time –
If you as my neighbor do not hear from me for 15 seconds, then bring the neighbourship
with me down.

Different EIGRP packets:

 Hello
 Update
 Query
 Reply
 ACK (Acknowledgement)

Hello packets are used for neighbor discovery. As soon as you send hello packets and receive
them your EIGRP routers will try to form the neighbor adjacency.

Update packets have routing information and are sent reliable to whatever router that
requires this information. Update packets can be sent to a single neighbor using unicast or to
a group of neighbors using multicast.

Query packets are used when your EIGRP router has lost information about a certain network
and doesn’t have any backup paths. What happens is that your router will send query packets
to its neighbors asking them if they have information about this particular network.

Reply packets are used in response to the query packets and are reliable.

ACK packets are used to acknowledge the receipt of update, query and reply packets. ACK
packets are sent by using unicast.

EIGRP Metric Calculation:

 Not all of the following metrics are used. Only bandwidth & delay. Look into the
default K values.

 B – Bandwidth
 D – Delay
 L – Load
 R – Reliability
 M – MTU

 By default K1=1, K2=0, K3=1, K4=0, K5=0.


 Reduced Metric Formula = (10,000,000 / Min BW + Sum of Interface Delays/10)
* 256.

EIGRP Feasibility Condition:

 An EIGRP route is a feasible successor route if its / advertised distance from our
neighbor is less than the feasible distance (FD) of the successor route.

Basic Configuration:

 Router ID – 32 bit value if not statically configured is the highest IPv4 address on
an up / up loopback interface or the non-loopback interface.

router eigrp 1 (AS number)


network 192.168.1.0 0.0.0.255 (/24)
network 192.168.2.0 0.0.0.255 (/24)
network 10.1.1.1 0.0.0.0 (/32)

network 0.0.0.0 (All my router interfaces to participate in EIGRP)

EIGRP Queries and Stuck In Active (SIA):

 EIGRP has a feasibility successor route in an event that the successor route goes
down.
 However, in some cases there does not a feasible successor route. What happens
when the successor route goes down then?
 Going active for a route – When an EIGRP speaking router sends one or more query
messages to its EIGRP speaking neighbors in an attempt to find an alternate path to
a network that is no longer available.
 Queries are further sent by the neighboring router to its EIGRP speaking neighbors
and so on.
 While this querying is going on, the router that initiated the query is ACTIVE state.
 If the router does not get a reply to its query within 3 minutes, then it considers that
router to be down. Let’s say one of the downstream query reply is dropped and the
initial router is stuck in active state.
 With changes, now, if in 90 seconds the router does not hear a reply for the query
that it sent, it sends a SIA query. If it receives the reply for SIA query, then the
router does not break the adjacency.

EIGRP Stub Routing:

 One way to limit the scope of SIA queries is by using stub routing.
 Queries not sent from non-stub routers to stub routers.
 Stub router is not going to advertise routes from one EIGRP neighbor to another
EIGRP neighbor.
 For example, if router A learns about network B from Router B, it does not advertise
that to Router C and vice-versa.
 Config: eigrp stub [option]

 Options:
o Connected – Stub router advertises connected routes matched with a network
command.
o Summary – Stub router advertises summarized routes
o Static – The stub router advertises statically configured routes, if the
redistribute static command has been configured.
o Leak-map name - The stub router’s dynamic prefixes are based on a leak-
map.
o Redistributed – The stub router advertises any redistributed routes
o Receive-only – Stub router does not advertise any routes

EIGRP Passive Interface:

 Not sending EIGRP HELLO messages out of that interface.


 Not going to form EIGRP neighbourship over that interface.
 Ignores any EIGRP messages received.

Two ways to advertise a subnet into EIGRP:

Using the network command, HELLO packets are sent out the connected interfaces that are
matched with it. It may include those interfaces where neighbourship is not required but
only that subnet should be advertised. By making the interface as PASSIVE, it doesn’t send
any EIGRP HELLOs.

Using the redistribute command. For example, redistribute connected command


redistributes all connected routes. But much less preferable due to the AD value.

EIGRP Summary Routes:

 For example, 172.16.3.0/24, 172.16.2.0/26, 172.16.1.0/24, 172.16.0.0/24, they


could all be summarized into 172.16.0.0./22

Variance:

++ To enable unequal cost load-balancing.


++ RIP and OSPF can do load-balancing only when they have multiple equal cost routes.
++ By default the routing table has only the Successor route.
++ By using variance, we could make the routing table include both Successor and Feasible
Successor routes.
++ Formula: FD of the Feasible Successor < Feasible Distance of the Successor *
Multiplier.
OSPF Fundamentals:

 Open Standard.
 Link State and is NOT distance vector.
 Establishes adjacencies with other routers. (Adjacency vs Neighbourship)
 Sends LSAs to other routers in an area.
 Constructs a Link State database from received LSAs.
 Runs the Dijkstra’s algorithm to determine the shortest path to a network.
 Attempt to inject the best path for each network into a router’s IP routing table.
 OSPF AD – 110
 Send HELLO packets to 224.0.0.5

Neighbourship vs Adjacency:

 Neighbors are routers that reside on the same network link.


 Exchange HELLO messages.
 Adjacencies are routers that are neighbors that exchange Link State Updates and
Database Description packets.

Parameters that need to match:

 HELLO Interval
 DEAD Interval
 Area ID
 Subnet Mask
 Stub Area Flag
 Authentication

Type of OSPF packets:

 HELLO
 Database Description (DBD)
 Link State Request
 Link State Update
 Link State Acknowledgement

Forming an adjacency:

Down
Attempt - only when statically defined a neighbor. E.g. NBMA. Trying to send HELLO to that
neighbor but not getting any HELLO from him.
Init – Sending OSPF HELLOs to 224.0.0.5 but still doesn’t see its own router ID
2-Way – Now sees itself in the OSPF HELLO sent by the neighbor
Exstart – Select a DR / BDR if ethernet network. Select Master / Slave.
Exchange – Exchange Database Descriptor (DBD) packets between each other. DBD has
the list of LSAs known to the router that sends it.
Loading – Sending Link State Request (LSR) and Link State Update (LSU) in return.
Full – OSPF adjacency fully established.

OSPF Areas:

 Area 0 – Backbone Area


 If you have multiple areas, they must be connected to the backbone area – Area 0.
 Routers bordering or connecting two different areas – Area Border Routers.

Configuration:

router ospf 1
network 192.168.1.0 0.0.0.255 area 1
network 1.1.1.1 0.0.0.0.0 area 1
network 10.0.0.0 0.0.0.255 area 1

OSPF Network Types:

 Broadcast Network – Multiple routers connected to a broadcast medium like a


switch. For example, let us say 6 OSPF running routers are connected to a switch,
each of them will have same OSPF database. Don’t need a full mesh. A DR and BDR
will be elected. Default HELLO interval – 10 seconds. HELLO interval x 4 = Dead
Interval.
 Point-to-Point Network – Only 2 routers connected to each other. No need for DR
or BDR. HELLO Interval – 10 seconds. Dead Timer = 40 seconds.
 NBMA Network – Example is Frame Relay. Let us say R1 is connected to R2, R3
and R4 via Frame-Relay. Cannot use multicast to discover a OSPF neighbor. Have to
statically define a neighbor. Need a DR and BDR. HELLO interval is 30 seconds and
DEAD timer is 120 seconds. Need to use a neighbor command.
 Point-to-Multipoint – No DR and BDR election. HELLO interval is 30 seconds and
DEAD timer is 120 seconds.

DR and BDR:

 Number of Adjacencies = [n * (n-1)]/2, where n is the number of routers.


 Full mesh needs a lot of adjacencies when there are more routers running OSPF.
 Solution – Elect a DR and BDR.
 224.0.0.6 – All designated and backup designated routers.

How a DR is elected?

 OSPF HELLOs are used to elect a DR.


 During a DR election, the router with the highest OSPF priority value wins.
 OSPF priority value is associated with an interface and can be a value in the range 0-
255.
 OSPF priority value of 0 means that the router will not become the DR.
 If the priorities tie, the router with the highest router ID (RID) becomes the DR.
 A router’s RID can be configured in router configuration mode, with the command
router-id id.
 If no RID is configured, highest loopback IP, if not highest interface IP becomes the
RID.

OSPF Timers:

 HELLO Timer – How often I send a OSPF HELLO


 DEAD Timer – The time that I will wait to receive a HELLO before tearing the
adjacency down.
DEAD – 4 x HELLO.

OSPF Passive Interface:

 Similar to EIGRP passive interface.


 Advertise a network over OSPF. But do not send OSPF HELLO packets over that
interface.
 Security purposes.
LSA Types:

 Routers belonging to an area should have an identical LSDB.


 Type 1 LSA – Router LSA. LSAs generated by routers within the area same area.
Advertising the directly connected links.
 Type 2 LSA – Network LSA. LSAs generated by the DR’s. Again within the area.
 Type 3 LSA – Summary LSA. Not necessarily advertising a summary route. Sent
by ABRs that has the information about link states, prefixes etc within that area.
 Type 4 LSA – Let us say we have a EIGRP running router that is redistributing
its routes into the OSPF areas via an ASBR. Type 4 LSAs are also originated by
the ABR and the destination that they advertise is the ASBR.
 Type 5 LSA – LSAs that tell about the external network. Sent by the ASBRs.
Advertise a destination external to the OSPF autonomous system.
 Type 6 LSA – Group membership LSA. Rarely used.
 Type 8 LSA – Link local LSA used with OSPF V3.
 Type 9, 10, 11 LSA – Reserved for future use.
 Stub Area – We do not want to see externally injected routes into the stub area.
Therefore, no Type-4 or Type-5 LSA. How to get to those external routes –
Default route to get to any network that is not being explicitly advertised.
 Totally Stubby Area – Not only will not include Type-3 LSA, Type-4 LSA or
Type-5 LSA, just have a single Type-3 LSA i.e. to send traffic to any destination,
send it here. Will not even receive information from the neighboring area.
 Not-So-Stubby Area – Type 7 LSA. External routes from another autonomous
system. Sent by ASBR. External routes cannot be injected to stub and totally
stubby areas which means that routers within these areas cannot do
redistribution. Routers in NSSA have the freedom to become ASBRs. Type 3 LSAs
are still advertised.
 Totally Not-So-Stubby Area – Similar to NSSA, but no Type-3 LSAs as well.

Router LSA:

1. If router has an OSPF speaking interface connected to a link on which DR is not


elected – stub network. Type 1 LSA is generated.
2. If an interface that has DR elected – transit network.
3. DR is not elected, but the neighbor is reachable.

 Each router creates a Type 1 LSA for itself and floods it throughout the same
area.

Network LSA:

 Type 2 LSA is generated for multi-access networks. To properly map all the
connected routers to a single multi-access network like LAN.
Summary LSA:

 Not advertising Type 1 and 2 LSAs across areas saves memory and reduces the
complexity for each run of the SPF algorithm.
 Type 3 does not contain all the detailed topology information.
 Type 3 LSA consists of each subnet and cost to reach that subnet from that ABR.
 Summarizes the information from Type 1 and Type 2 LSAs.

OSPF metric calculation:

 Cost = Reference Bandwidth / Interface Bandwidth

OSPF Database Exchange without DR:

 After the 2-way state, next state is ExStart.

When no DR exist, the next step is that neighbors want to start with LSDB exchange. They
use the following general process:

1. Discover the LSAs known to the neighbor but unknown to me.


2. Discover the LSAs known by both the routers but the neighbors LSA is more up to
date.
3. Ask the neighbor for a copy of all the LSAs identified in the first two steps.

 The router goes into ExStart state as soon as the DBD message is received.
 Election is held for Master and Slave.
 Router with higher RID is elected as master.
 DD packets are sent using multicast IP address 224.0.0.5.
 After Master / Salve election, the neighbor transitions into the Exchange state.

OSPF Database Exchange with DR:

 To exchange database with DR/BDR, communication with DR/BDR is done using a


All-DR multicast address – 224.0.0.6.
 DR performs the database exchange with the same messages but sends the
messages to 224.0.0.5.

Periodic Flooding:

 OSPF re-floods each LSA every 30 minutes based on each LSAs variable.
 The router that creates the LSA sets this age to 0 seconds.
 When no changes to the LSA occur for 30 minutes, the owning router increments
the sequence number, resets the timer to 0 and re-floods the LSA.

OSPF E1 vs E2 routes:

 E routes are external to the OSPF autonomous system.


 E2 means OSPF external Type 2
 E1 means OSPF external Type 1

 When you redistribute a route say EIGRP / RIP etc. into OSPF, by default it is
Type 2 route with a cost of 20. It creates a Type 5 LSA. Type 2 route says don’t
change the cost that am advertising to you. While Type 1 could take the cost
from the one that ASBR advertises and you could calculate the cost that is
needed to reach the ASBR as well.

OSPF Route Filtering:

 OSPF cannot have different link state databases within an area. Cannot filter
selective routes for OSPF database.
 Just because OSPF knows the route, there is no guarantee that it will be installed
into the IP routing table.
 Using distribute-list we can block the route to be injected into the IP routing table
but not altering the LSDB.
 We can configure filter-lists to filter LSAs. Create a prefix-list and apply it as a
filter-list under the OSPF process.

OSPF Route Summarization:

 Could do summarization in the ABR router using area range command.


 Could do summarization in the ASBR using summary-address command.

OSPF Virtual Links:

 Let us say Area 0 is connected to Area 1 to Area 2.


 Indirectly connect Area 2 to Area 0 as if it was connected directly.
 Virtual Links to help achieve this.
BGP Fundamentals:

 BGP is an exterior gateway routing protocol that routes between AS.


 Forms neighbourship.
 Neighbor’s IP address is explicitly configured.
 TCP session is established between neighbors.
 Advertises address prefix and length (called Network Layer Reachability
Information).
 Advertises a collection of PATH attributes that can be used for PATH selection.
 Path Vector routing protocol.

router bgp <AS number>


bgp router-id 1.1.1.1
network 192.0.2.0
neighbor 192.0.2.2 remote-as 65001
neighbor 10.1.1.2 next-hop-self

 iBGP – BGP neighbor in the same AS.


 eBGP – BGP neighbor in different AS.
 Within an iBGP system, next-hop IP does not change during route
advertisements.

BGP Data Structures:

 BGP neighbor table


 BGP table or BGP RIB
 IP routing table

BGP Message Types:

Forming a BGP neighbourship:

 BGP requires we statically define the IP of neighbor.


 IDLE state
 Connect state – TCP session is setup
 Open Sent – R1 sends an OPEN message to R2 about itself
 Open Confirm – R2 sends an OPEN message back to R1
 Established – Keep-alives exchanged

Open Message – BGP version number, Local AS number, Hold Time, BGP Router ID,
Optional parameters.
Keepalive Message – A message header that keeps the HOLD timer from expiring.
Update Message – Can contain withdrawn routes, path attributes and NLRI.
Notification Message – Contains an error code, error subcode information about the error.

BGP Peer Group:

 Reduces a BGP router’s CPU load by creating an update once per group, rather
than once per neighbor.

Вам также может понравиться