BGP: Border Gateway Protocol Basics
BGP (Border Gateway Protocol) is the routing protocol that makes the internet work. It's the protocol that connects different networks (autonomous systems) together and determines how data flows across the global internet as defined in RFC 4271. This comprehensive guide explains BGP, how it works, and why it's critical to internet infrastructure.
What is BGP?
BGP is an exterior gateway protocol (EGP) designed to exchange routing information between different autonomous systems (AS). It's the only protocol used for routing between ISPs and is responsible for maintaining the internet's routing table.
Key Characteristics
Protocol type:
Classification: Path vector protocol
Layer: Application layer (TCP port 179)
Purpose: Inter-domain routing
Scale: Internet-scale (millions of routes)
Core features:
Policy-based routing
Path selection control
Loop prevention
Scalability
Stability over speed
Incremental updates
BGP vs IGPs: ``` BGP: Between autonomous systems IGPs (OSPF, EIGRP): Within autonomous systems
BGP: Policy-driven IGPs: Metric-driven
BGP: Slow convergence, stable IGPs: Fast convergence ```
Autonomous Systems (AS)
What is an AS?
Definition:
Collection of IP networks under single administrative control
Unique AS number (ASN)
Common routing policy
Single routing protocol
AS Number (ASN): ``` 16-bit: 1 - 65535 (original) 32-bit: 1 - 4294967295 (extended)
Reserved ranges: 0: Reserved 64512-65534: Private use 65535: Reserved 23456: AS_TRANS (transition) ```
Types of AS:
Stub AS:
Single connection to internet
One upstream provider
No transit traffic
Most common type
Multi-homed AS:
Multiple connections to internet
Multiple providers
Redundancy
No transit traffic
Transit AS:
Provides transit for other AS
ISP, backbone provider
Carries third-party traffic
Complex routing policies
AS Examples
Major AS numbers:
AS7018: AT&T
AS3356: Level 3 (Lumen)
AS1299: Telia
AS174: Cogent
AS6939: Hurricane Electric
AS15169: Google
AS8075: Microsoft
AS16509: Amazon
BGP Types
eBGP (External BGP)
Between different AS:
Connects different organizations
Different AS numbers
TTL = 1 (directly connected)
Next-hop changes
Configuration example (Cisco):
router bgp 65001
neighbor 203.0.113.1 remote-as 65002
neighbor 203.0.113.1 ebgp-multihop 2
Characteristics:
Administrative distance: 20
Next-hop: Usually changed
AS-PATH: Incremented
Used for: Internet routing
iBGP (Internal BGP)
Within same AS:
Distributes external routes internally
Same AS number
Full mesh or route reflectors
Next-hop preserved
Configuration example (Cisco):
router bgp 65001
neighbor 10.0.0.2 remote-as 65001
neighbor 10.0.0.2 update-source Loopback0
neighbor 10.0.0.2 next-hop-self
Characteristics:
Administrative distance: 200
Next-hop: Preserved (unless next-hop-self)
AS-PATH: Not incremented
Used for: Internal route distribution
iBGP requirements:
Full mesh (all routers peer with all)
Or route reflectors
Or confederation
Prevents loops
BGP Attributes
Path Attributes
Well-known mandatory:
AS_PATH:
List of AS numbers traversed
Loop prevention mechanism
Shorter path preferred
Primary path selection criterion
NEXT_HOP:
IP address of next hop router
Changed by eBGP
Preserved by iBGP (usually)
Must be reachable
ORIGIN:
How route was injected into BGP
IGP (i): network command
EGP (e): EGP (obsolete)
Incomplete (?): redistribution
Well-known discretionary:
LOCAL_PREF:
Preference within AS
Higher value preferred
iBGP only
Default: 100
ATOMIC_AGGREGATE:
Route was aggregated
Information lost
Informational
Optional transitive:
AGGREGATOR:
AS and router that aggregated
Informational
Passed to other AS
COMMUNITY:
Route tagging
Policy implementation
32-bit value
Well-known communities
Optional non-transitive:
MED (Multi-Exit Discriminator):
Suggests preferred entry point
Lower value preferred
Compared between routes from same AS
Not passed to other AS
ORIGINATOR_ID:
Route reflector attribute
Original router ID
Loop prevention
CLUSTER_LIST:
Route reflector clusters
Loop prevention
BGP Path Selection
Decision Process
Order of preference:
1. Highest Weight (Cisco-specific):
Local to router
Not advertised
Default: 0
Higher preferred
2. Highest LOCAL_PREF:
Within AS
Default: 100
Higher preferred
3. Locally originated:
Prefer routes originated by this router
network command > redistribute
4. Shortest AS_PATH:
Fewer AS hops
Primary internet routing criterion
Can be manipulated (AS prepending)
5. Lowest ORIGIN:
IGP (i) > EGP (e) > Incomplete (?)
Rarely used for selection
6. Lowest MED:
Between routes from same AS
Suggests preferred entry
Lower preferred
7. eBGP over iBGP:
External routes preferred
Administrative distance: eBGP 20, iBGP 200
8. Lowest IGP metric to NEXT_HOP:
Shortest internal path
Hot potato routing
9. Oldest route:
Stability
Prefer established paths
10. Lowest router ID:
Tiebreaker
Deterministic
Path Selection Example
Scenario: ``` Router receives 3 routes to 192.0.2.0/24:
Route A: AS_PATH: 65002 65003 LOCAL_PREF: 100 MED: 50
Route B: AS_PATH: 65004 LOCAL_PREF: 150 MED: 100
Route C: AS_PATH: 65005 65006 65007 LOCAL_PREF: 100 MED: 10
Selection process: 1. Weight: All equal (default 0) 2. LOCAL_PREF: Route B wins (150 > 100)
Route B selected ```
BGP Messages
Message Types
OPEN:
Establishes BGP session
Negotiates parameters
AS number, BGP version, hold time
Capabilities
UPDATE:
Advertises new routes
Withdraws old routes
Path attributes
NLRI (Network Layer Reachability Information)
KEEPALIVE:
Maintains session
Sent periodically
Default: 60 seconds
Prevents timeout
NOTIFICATION:
Error condition
Closes session
Error code and subcode
Debugging information
BGP Session States
Idle:
Initial state
Waiting to start
Connect:
Waiting for TCP connection
Active:
TCP connection failed
Trying to reconnect
OpenSent:
OPEN message sent
Waiting for OPEN
OpenConfirm:
OPEN received
Waiting for KEEPALIVE
Established:
Session active
Exchanging routes
Normal operation
State transitions:
Idle → Connect → OpenSent → OpenConfirm → Established
↓ ↓ ↓ ↓
Active ←────┴───────────┴──────────────┘
BGP Configuration
Basic eBGP Configuration
Cisco IOS:
router bgp 65001
bgp router-id 1.1.1.1
neighbor 203.0.113.1 remote-as 65002
neighbor 203.0.113.1 description ISP_A
!
address-family ipv4
network 192.0.2.0 mask 255.255.255.0
neighbor 203.0.113.1 activate
exit-address-family
Juniper:
set protocols bgp group external type external
set protocols bgp group external peer-as 65002
set protocols bgp group external neighbor 203.0.113.1
set policy-options policy-statement export-bgp term 1 from protocol static
set policy-options policy-statement export-bgp term 1 then accept
Basic iBGP Configuration
Cisco IOS:
router bgp 65001
bgp router-id 1.1.1.1
neighbor 10.0.0.2 remote-as 65001
neighbor 10.0.0.2 update-source Loopback0
neighbor 10.0.0.2 next-hop-self
!
address-family ipv4
neighbor 10.0.0.2 activate
exit-address-family
Route Reflectors
Purpose: Avoid iBGP full mesh
Configuration:
router bgp 65001
neighbor 10.0.0.2 remote-as 65001
neighbor 10.0.0.2 route-reflector-client
Hierarchy: ``` Route Reflector (RR) ├── Client 1 ├── Client 2 └── Client 3
Clients peer only with RR RR reflects routes between clients Reduces peering requirements ```
BGP Policies
Route Filtering
Prefix lists: ``` ip prefix-list ALLOW-CUSTOMER permit 192.0.2.0/24 ip prefix-list ALLOW-CUSTOMER deny 0.0.0.0/0 le 32
router bgp 65001 neighbor 203.0.113.1 prefix-list ALLOW-CUSTOMER out ```
AS-PATH filtering: ``` ip as-path access-list 1 permit ^65002$ ip as-path access-list 1 deny .*
router bgp 65001 neighbor 203.0.113.1 filter-list 1 in ```
Route Manipulation
AS-PATH prepending: ``` route-map PREPEND permit 10 set as-path prepend 65001 65001 65001
router bgp 65001 neighbor 203.0.113.1 route-map PREPEND out
Result: Makes path less attractive ```
LOCAL_PREF: ``` route-map SET-LOCAL-PREF permit 10 set local-preference 200
router bgp 65001 neighbor 10.0.0.2 route-map SET-LOCAL-PREF in
Result: Prefer this path ```
MED: ``` route-map SET-MED permit 10 set metric 50
router bgp 65001 neighbor 203.0.113.1 route-map SET-MED out
Result: Suggest this entry point ```
Communities
Well-known communities:
NO_EXPORT: Don't advertise to eBGP peers
NO_ADVERTISE: Don't advertise to any peer
LOCAL_AS: Don't advertise outside confederation
Custom communities:
65001:100 - Customer routes
65001:200 - Peer routes
65001:300 - Upstream routes
Configuration: ``` ip community-list 1 permit 65001:100
route-map SET-COMMUNITY permit 10 set community 65001:100
router bgp 65001 neighbor 203.0.113.1 send-community neighbor 203.0.113.1 route-map SET-COMMUNITY out ```
BGP Security
Common Threats
Route hijacking:
Malicious AS announces prefixes it doesn't own
Traffic redirected
Data interception
Service disruption
Route leaks:
Unintentional announcement of routes
Misconfiguration
Causes internet outages
Traffic blackholing
Prefix hijacking:
More specific prefix announced
Longest prefix match wins
Traffic diverted
Security Measures
Prefix filtering:
Filter customer announcements
Only accept owned prefixes
Bogon filtering (invalid ranges)
Maximum prefix limits
AS-PATH filtering:
Validate AS-PATH
Prevent private AS in public internet
Filter known bad AS
RPKI (Resource Public Key Infrastructure):
Cryptographic validation
ROA (Route Origin Authorization)
Validates prefix ownership
Prevents hijacking
BGPsec:
Path validation
Cryptographic signatures
Prevents path manipulation
Not widely deployed
Maximum prefix: ``` router bgp 65001 neighbor 203.0.113.1 maximum-prefix 1000
Prevents route table overflow Detects misconfigurations ```
BGP in Practice
ISP Peering
Types of peering:
Transit:
Pay for connectivity
Full routes or default
Upstream provider
Peering:
Mutual exchange
No payment
Internet exchange points (IXP)
Bilateral agreements
Customer:
Provide transit
Receive payment
Announce customer routes
Multi-homing
Benefits:
Redundancy
Load balancing
Provider independence
Better performance
Configuration:
Two or more ISP connections
Own AS number
BGP with each provider
Policy-based routing
Inbound traffic control:
AS-PATH prepending
MED
Communities
More specific prefixes
Outbound traffic control:
LOCAL_PREF
AS-PATH
Weight
Default routes
Internet Exchange Points (IXP)
Purpose:
Neutral peering location
Multiple networks interconnect
Reduced costs
Better performance
Major IXPs:
DE-CIX (Frankfurt)
AMS-IX (Amsterdam)
LINX (London)
Equinix (multiple locations)
Troubleshooting BGP
Common Issues
Session not establishing:
Check: TCP connectivity (port 179)
Check: AS numbers correct
Check: IP addresses correct
Check: Firewall rules
Routes not received:
Check: BGP session established
Check: Route filters
Check: Maximum prefix limit
Check: Soft reconfiguration
Suboptimal routing:
Check: Path attributes
Check: Policies
Check: AS-PATH
Check: LOCAL_PREF
Diagnostic Commands
Cisco IOS:
show ip bgp summary
show ip bgp neighbors
show ip bgp
show ip bgp 192.0.2.0
show ip route bgp
debug ip bgp updates
Verification: ``` show ip bgp summary - Check session state (Established) - Check prefixes received/sent - Check uptime
show ip bgp neighbors 203.0.113.1 - Detailed neighbor information - Capabilities - Statistics ```
IPv6 BGP
MP-BGP (Multiprotocol BGP)
Configuration:
router bgp 65001
neighbor 2001:db8::1 remote-as 65002
!
address-family ipv6
neighbor 2001:db8::1 activate
network 2001:db8:1::/48
exit-address-family
Differences:
Same protocol, different address family
IPv6 next-hop
IPv6 prefixes
Otherwise identical to IPv4 BGP
Best Practices
Design
1. Use private AS for internal:
64512-65534 for internal use
Don't leak to internet
2. Implement route filtering:
Filter customer announcements
Bogon filtering
Maximum prefix limits
3. Use route reflectors:
Avoid full mesh
Hierarchical design
Redundant RRs
Operations
1. Monitor BGP:
Session state
Route counts
Flapping
Changes
2. Document policies:
Peering agreements
Route policies
Community usage
Contact information
3. Implement security:
RPKI validation
Prefix filtering
AS-PATH filtering
Authentication
Troubleshooting
1. Layer by layer:
Physical connectivity
IP connectivity
TCP session
BGP session
Route exchange
2. Verify configuration:
AS numbers
IP addresses
Route policies
Filters
3. Check logs:
Session flaps
Route changes
Errors
Notifications
Conclusion
BGP is the routing protocol that powers the internet, enabling different autonomous systems to exchange routing information and determine optimal paths. Understanding BGP is essential for anyone working with internet infrastructure, ISPs, or large enterprise networks.
Related Articles
Routing Fundamentals
- Routing - IP routing basics
- Default Gateway - First hop routing
- IPv4 CIDR Notation - Route aggregation
- Anycast - BGP-based routing
Internet Infrastructure
- Internet Service Providers - ISP operations
- CDN - Content delivery networks using BGP
- IPv6 Adoption - BGP and IPv6
- Autonomous Systems - AS numbers
Network Protocols
- ICMP - Error reporting in routing
- TCP/IP Model - Protocol stack
- DNS Servers - DNS and BGP
Security
- DDoS Attacks - BGP hijacking
- IP Reputation - BGP blackholing
Explore More
- Networking Basics - Essential concepts
- Protocols - Internet protocols hub
Key takeaways: - BGP connects autonomous systems - Path vector protocol - Policy-based routing - eBGP between AS, iBGP within AS - AS_PATH prevents loops - Path selection based on attributes - Scalable to internet size - Security critical (RPKI, filtering) - Used by ISPs and large enterprises - Complex but powerful
BGP's flexibility and scalability make it the only protocol capable of handling internet-scale routing, making it one of the most critical protocols for global internet connectivity.
