1) NBR are maynot be in same subnet

2 nbr maynot be directly connected to each other

3) Unlike IGP protocol, BGP advertise the whole subnet, instead of advertising hte
individual network
4) network command has diff purpose in BGP in comparision to IGP protocols

BGP states
1) IDLE -
once you config BGP on it it will strt intitiating TCp connection with BGP config
peer and start listening TCP connection from peer

if any error occur at any stage, BGp session is terminated and returned into IDle.

***************************************
* TCP 179 not open
* AS no. configuration incorrect
* peer address config incorrect

2) connect -
completes 3 way hand shakes
after successful TCP connection establishment with the peer device sends open msg
to its peer

once open msg is sent it jumps to open sent state

3) Active state -
if rtr not able to estb a successful TCP session then it ends up in Active state

* it tried to restart another TCP session with the peer, if that is succesful then
is sends an open msg to the peer and jumps to opensent state

4) opensent state
* RTR sends an open msg and receives and open msg from its peer
once it receives and open msg it checks it's validity Like ( BGP Ver No., AS no,
Nbr statement )

5) if everything is ok, then it sends and keep alive msg and jumps to openconfirm
state

BGP messages
* All BGP msgs are unicast to its peer over TCP connection
* TCP is responsible for retransmission, fragmentation, and ACkmnt

BGP mgs type

1) OPen
2) Keep alive
3) Update
4) Notification
5) route refresh

1) OPen msg - RTR uses this msgs to identify itself and validate the attributes of
its peer
open msg is always sent after the TCP connection estblised
* version, AS, Hold time, BGP identifier, opetional parameter lenght, opentional
parameter

2) keep Alive - if router accepts the abpove parameter of its peer then it sends a
keep alive msg
* After connection estb, BGP sends Keep alive messages after every 60 seconds.
* after missing 3 keep alive msgs from peer BGP goes down

3) Update - RTR advertise the NLRI ( network layer reachability information, NLRI
is ad of new prefixes )

* it sends and update msgs after every change, wheather, to advertise the
router the route ot for withdran parameter.
unfeasable route lenght, withdrawn routes, Total path attribution lenght, path
attributes,, network layer reachability information

4) Notification ::- sent when some error happened

* error || error subcode || error Data

5) route refresh - The BGP (Border Gateway Protocol) Route Refresh message is used
to update BGP routers about changes in routing information without tearing down the
entire BGP session. This message is particularly useful when changes occur in the
network, such as new routes being added or existing routes being withdrawn.

Two Types: There are two types of BGP Route Refresh messages: "soft" and "hard."
The "soft" option allows a BGP speaker to request updates for a specific address
family, while the "hard" option requests updates for all address families.

BGP best path selection order

1) weight( highest)
2) Locally significant cisco propierty attribute
3) Locally prefernce ( Highest
4) Locally originated
5) AS- path( shortest
6) Origin
7) MED ( Lowest
8) EBGP Larned routes over IBGP learned routes
9) Smallest IGP metric to next hop value

1) Next Hop – It is used to list the next-hop IP address that can be used to reach
the prefix. The router can’t use this router if no route is available to reach the
next-hop.
2) Weight – It is used to give advantage to the path with the highest weight. It is
locally significant to the router and is Cisco proprietary.
3) Local Preference – It is used to give advantage to the path with highest local
preference. It is used within an AS and exchanged between iBGP routers. Default is
set to ***100***.
4) Locally injected routes – Routes that are injected using network command are
called Locally injected routes. These routes are better than iBGP/eBGP learned.
5) AS Path – It is applied to outbound routes. It helps in deciding the best
inbound path. With the AS-Path attribute, two things can be accomplished that are:
prepend or filter. Smaller is preferred.
6) Origin – This attribute is used to identify the originating source of the route.
The origin codes are as follows (listed in order of preference for route
selection):
7) i (IGP) – It is emerged from an IGP (interior gateway protocol), such as OSPF.
This usually indicates that network command under the BGP process was used to
inject the route into BGP. An origin code that has “i” is most desired.
8) e (EGP) – It is originated from an EGP (external gateway protocol).
? (incomplete) – It means the origin of it is Unknown. This specifies that the
route was redistributed into BGP from either connected, static, or IGP routes. An
origin code that has “?” is the least desired.
9) Multi-Exit Discriminator (MED) – This attribute is applied to the outbound
routes. It finds the best inbound path into the AS. When viewing the BGP routing
table, MED is seen as the BGP metric.
By default, the MED value is 0 and a smaller/lower value is preferred.

Neighbour type – In this, the eBGP is preferred over iBGP

IGP metric – With this attribute, the route with nearest IGP neighbour (lowest IGP
metric) is preferred.
eBGP route – The Oldest route is preferred.
Neighbour Router ID – Lowest is preferred.
Neighbour IP Address – Lowest is preferred

***** other tie breaking check occur if no best path

oldest route, lowest R-ID, lowest interface IP, Etc..

BGP commands
1) Show ip BGP ( shows all BGP details
2) ip bgp 1.1.1.1 soft in ( to soft reset BGP )
3) traceroute x.x.x.x source loopback 0 ( to trace route the BGP )

Manipulating BGP best path selection

vector attributes can be manually modified to define differnet policy for

different routes,

Attributes typically modified are

- Weight
- Local preference
- AS path
- MED

**Inbound routing policy affect outbound traffic

change weight or local-pref into affect traffic out

**Outbound routing policy affects incoming traffic

change AS-path or MED to affect traffic in
IN BGP, what is the order of preference?

The following is the order of preference for inbound updates:

Route-map
Filter-list
Prefix-list
distribute-list
The following is the order of preference for outgoing updates:

Filter-list
Route-map | unsuppressed-map
Advertise-map (conditional-advertisement)
Prefix-list
distribute-list.

trouble shooting :: -

1) intermittent issue :: -
1.1) first we will ping / trace to exact where the drops are happening.
1.2) Neighbor - sh ip bgp summary ( is bgp up / down or stable ) if stable then
Route - sh ip bgp ( is route are stable in bgp database ) if routes are not
available because of ACL, prefix list....etc.
1.3) we can also check timers if timers are not matching Like keep alive / whole
timers are not correct
1.4) Route selection - route map - to influence the incoming/ outgoing traffic.
1.5) Routing Loop :: - No routing loops are there.
1.6) network conjunction :: - due to traffic conjunction ( tool - SNMP, net flow )
1.7) Any hardware / Software failure is there or not. CPU utilization etc...

2) merge two companies / BGP peering b2n two network.

2.1) Customer types in BGP. ibgp, ebgp, policies, routing policies, AS No,
2.2) Configure and verify there is no issue, Ping / trace
2.3)