What is a BGP session?

Understand BGP sessions (Border Gateway Protocol): peering, ASN, eBGP/iBGP, RPKI security, and how to open a BGP session with HolyCloud.

~14 min read Beginner #bgp session #bgp #asn #ebgp #ibgp #routing #peering #rpki #datacenter

What is a BGP session?

On the Internet, data packets travel across a complex mesh of routers and autonomous systems to reach their destination. This global routing relies on routing protocols, and at the heart of that system is the Border Gateway Protocol (BGP).

A BGP session is the connection established between two routers (peers) to exchange routing information. Thanks to these sessions, networks advertise IP prefixes to each other and traffic finds a viable path from one end of the Internet to the other.

In short: BGP is the “postal service” of the Internet. A BGP session is the secure conversation between two post offices (routers) agreeing on which paths to use.

What is BGP?

BGP (Border Gateway Protocol) lets the Internet operate as interconnected networks. Unlike interior gateway protocols (IGPs) that handle routing inside a single autonomous system (AS), BGP works between ASes.

BGP drives IPv4 and IPv6 routing. Without it, the Internet would be a collection of isolated networks unable to communicate efficiently at global scale.

BGP session and peering

For BGP to work, two routers must:

Open a TCP connection (port 179)
Authenticate (often MD5 / TCP-AO)
Exchange routing tables (prefix advertisements)
Keep the session alive (KEEPALIVE messages)

This relationship is called peering. Each BGP neighbor is a peer. Attributes on each route (AS_PATH, LOCAL_PREF, MED, COMMUNITIES…) then guide best-path selection.

Why BGP matters for the Internet

BGP is critical because routers can exchange reachability information even on highly complex topologies. It provides:

Global reachability

Any correctly peered network can reach (and be reached by) the rest of the Internet.

Network stability

By continuously updating routes, BGP adapts to failures, congestion, and topology changes.

Traffic optimization

Operators influence flow (multi-homing, traffic engineering, transit preference) through BGP attributes and policies.

Key functions of a BGP session

Routes and routing tables

A BGP route describes how to reach a prefix (e.g. 203.0.113.0/24). Routes live in the RIB, then the selected paths enter the forwarding table.

BGP session (peering)

Without an Established session, no routes are exchanged. Typical states: Idle → Connect → OpenSent → OpenConfirm → Established.

Path selection

When multiple routes exist to the same destination, BGP’s decision process (LOCAL_PREF, AS_PATH, ORIGIN, MED, etc.) picks the best path.

How BGP works

Peering explained

Two routers agree to exchange routes, open TCP, authenticate, then sync tables. Each BGP attribute influences selection.

Autonomous Systems (AS)

An AS is a set of networks under one administrative policy, identified by an ASN (e.g. AS212477). ISPs, cloud providers, and large enterprises each run their own AS.

Advertising and filtering

Peers advertise reachable prefixes. On receive, you filter (prefix-lists, as-path filters, RPKI) to accept only legitimate announcements — essential against hijacks and leaks.

[Your AS / router]  ---- eBGP session ----  [HolyCloud edge]  ----> Transit / Internet peering
         │                                        │
    Your prefix ads                         Filtering + RPKI

BGP configuration types

Internal BGP (iBGP)

Sessions inside the same AS. They sync the internal view. iBGP routes are generally not re-advertised to other iBGP peers (loop prevention), hence full-mesh, route reflectors, or confederations.

External BGP (eBGP)

Sessions between different ASes. This underpins Internet routing and most customer BGP sessions with a hosting or transit provider.

iBGP vs eBGP

-----------	------	------
Hop / TTL	Often internal multi-hop	Often adjacent (TTL 1)
Typical use	Operator core	Internet / customer interconnect

Benefits of a BGP session

Scalability

BGP handles tens/hundreds of thousands of prefixes — suitable for ISPs and large networks.

Control and reliability

Granular policies, multiple paths, automatic failover if a transit fails.

Traffic management

Communities, AS-path prepend, LOCAL_PREF: tools to balance or prefer links.

Challenges and security best practices

Preventing BGP hijacking

An attacker advertises a prefix they do not own. Mitigations:

Strict filtering (only authorized prefixes)
RPKI / ROA origin validation
Continuous route monitoring

Limiting route leaks

Accidentally re-advertising learned routes can disrupt the Internet. Use aggregation, communities (no-export), and clear export policies.

Securing the session itself

MD5 / TCP-AO authentication
ACLs limiting sources on port 179
Regular audits of prefix-lists and peers

Common use cases

Multi-homing

Connect to multiple providers for redundancy. BGP selects (or balances) paths.

Load distribution

Influence inbound/outbound traffic across transit or peering links.

Announcing your own prefixes (BYOIP / ASN)

With an ASN and PI/PA prefixes, a BGP session at HolyCloud announces them to the Internet with anti-DDoS and dual-stack.

BGP session at HolyCloud

HolyCloud offers BGP sessions Dual Stack (IPv4 + IPv6) or IPv6 Only, with:

eBGP peering to our ASN
GRE / VXLAN support depending on design
Anti-DDoS protection
Fast delivery and support

See BGP Session offers →

For deeper technical parameters (peer IP, prefix-lists, communities), see BGP concepts for customers.

FAQ — BGP session