What is BGP? Route Advertisements and the Backbone of the Internet
08 May 2026
Border Gateway Protocol (BGP) is the foundational protocol that connects tens of thousands of independent networks and routes data traffic across the internet. In this guide, we examine how BGP works, route announcements, security vulnerabilities, and the physical internet backbone from a technical perspective.
October 4, 2021. Billions of users worldwide suddenly lost access to Facebook, WhatsApp, and Instagram for several hours. Stock prices wavered, small businesses lost contact with their customers, and Facebook employees were even locked out of their own buildings. The cause was not a cyberattack. It was a small configuration error in a protocol most users have never heard of: BGP.
How Does the Internet Actually Work?
It is a common misconception to think of the internet as one massive, unified computer network. In reality, the internet is a vast structure formed by the interconnection of tens of thousands of distinct networks — those of telecom operators, major technology companies, universities, and government agencies.
Each of these independent networks is called an Autonomous System (AS) and carries a unique identifier known as an ASN. When you send an email, your message is broken into small data packets. For these packets to travel from your device to the destination server, they typically need to pass through multiple Autonomous Systems. So who decides which path they take? This is precisely where BGP comes in.
What Is the BGP?
Border Gateway Protocol (BGP) is the routing protocol that enables Autonomous Systems on the internet to communicate with one another and find the optimal path for data packets to reach their destination. In technical terms, it is classified as a Path Vector protocol.
It is more accurate to think of BGP not as a routing protocol, but as a policy protocol. Its primary function is not to find the shortest path, but to select the most appropriate route based on policies defined by network administrators.
An analogy: Imagine sending a physical package from Istanbul to New York. Your postal service needs agreements with other countries’ postal networks and must know which transit hubs the package will pass through. Sometimes the shortest route is not the most economical; sometimes a specific country is avoided for political reasons; sometimes trusted partners are preferred. BGP makes exactly these kinds of decisions for internet traffic — in seconds.
How Do Route Announcements Work?
At the heart of BGP’s operating logic are Route Announcements. Each network announces to its neighboring networks which IP address blocks it can reach and how.
Think of it as a vast highway system: at every junction, there are signs indicating where each road leads. BGP routers constantly read and update these signs. For example:
- A network tells its neighbors: “You can reach Google’s IP blocks through me.”
- Another network announces: “Route traffic destined for users in Türkiye to me — I’ll forward it.”
As these announcements propagate across thousands of networks, routers in every corner of the world build their own routing maps.
Technical Details: How Does BGP Make Decisions?
Core Architecture
BGP uses TCP port 179 for reliable data transfer. Neighbor relationships (peers) are not discovered automatically — they must be configured manually. This is a fundamental indicator of why BGP is a policy protocol: who communicates with whom, and how, is entirely under the control of network administrators.
iBGP vs. eBGP
BGP is divided into two types based on where it operates:
- eBGP (External BGP): Handles routing between different autonomous systems. This is the core structure that forms the internet’s backbone.
- iBGP (Internal BGP): Carries BGP information between routers within the same autonomous system. Due to the Split Horizon rule in iBGP, information received from one neighbor is not forwarded to another by default — to overcome this, a Route Reflector or Full Mesh topology is used.
Best Path Selection Algorithm
Unlike interior routing protocols such as OSPF or EIGRP, BGP does not select the “shortest path” — it selects the “most preferred path.” When a router learns multiple paths to the same destination, it evaluates the following criteria in order, stopping at the first match:
- Weight — Higher wins (Cisco-specific, local router only)
- Local Preference — Higher wins; determines exit point within the AS
- Locally Originated — Whether the route was originated by this router
- AS_Path Length — Fewer AS hops wins
- Origin Type — IGP preferred over EGP or Incomplete
- MED (Multi-Exit Discriminator) — Determines entry point into neighboring AS; lower wins
- eBGP over iBGP — Routes learned via eBGP are preferred
- IGP Metric — Lower IGP cost to the BGP next-hop wins
Convergence: BGP Is Actually Slow
One of BGP’s most misunderstood characteristics is its speed. Contrary to popular belief, BGP is not a fast protocol that makes decisions in milliseconds. Propagating a route change across the entire world can take minutes. This slowness is a deliberate design choice: BGP prioritizes stability over speed, so that minor fluctuations do not destabilize the entire system.
Scale and Looking Glass
As of 2025, the global BGP routing table contains more than 950,000 routes — and this number grows every year. Link-state algorithms like OSPF cannot realistically manage a database of this scale. BGP keeps the internet scalable by propagating only incremental changes.
Looking Glass servers are publicly accessible tools that allow you to view BGP tables from different vantage points around the world. They are among the first resources network engineers turn to when troubleshooting.
BGP Security Threats: Hijacking and Leaks
BGP is fundamentally a trust-based protocol: when a network announces “this IP block belongs to me,” its neighbors accept this by default. This trust model opens the door to two distinct threats.
- BGP Hijacking (Route Hijacking): An AS deliberately announces an IP block that does not belong to it. This can result in traffic being misdirected, intercepted, or dropped entirely.
- BGP Leak (Route Leak): Typically caused by a misconfiguration, this occurs when a legitimate route is propagated to a wider audience than intended. Even without malicious intent, the impact can be devastating.
To counter these threats, cryptographic validation mechanisms such as RPKI (Resource Public Key Infrastructure) and BGPsec have gained traction in recent years — though global adoption is not yet complete.
The Physical Backbone of the Internet
When we talk about cloud technology, it is easy to imagine data floating through the air. In reality, the internet has a very concrete, physical backbone: a massive network of fiber optic cables connecting data centers, countries, and continents.
Submarine fiber optic cables laid across ocean floors carry approximately 99% of intercontinental internet traffic. Tier 1 carriers — AT&T, Verizon, Lumen, Telia, and others — own this infrastructure. They exchange traffic among themselves at no cost through peering arrangements, keeping the global internet flowing without interruption.
BGP announcements are the logical system that, in seconds, switches the routing of traffic flowing at the speed of light across these undersea fiber networks.
Case Study: The Facebook Outage of 2021
On October 4, 2021, Facebook engineers ran an erroneous command during routine maintenance. As a result, Facebook’s BGP routers sent a route withdrawal announcement to the internet — effectively signaling that all routes to Facebook’s IP addresses were no longer valid.
This announcement spread across the world in seconds, and the global internet erased Facebook’s IP addresses from its map. What made matters even more critical: Facebook’s internal systems, authentication services, and data center access cards all relied on the same infrastructure — meaning engineers could not fix the issue remotely. They had to physically travel to the data centers and intervene directly on the hardware.
This incident illustrated both the power and the fragility of BGP in a single moment: a single erroneous announcement can erase one of the world’s largest technology companies from the internet’s map for hours.
Frequently Asked Questions
What does the BGP protocol do?
BGP (Border Gateway Protocol) enables independent networks (Autonomous Systems) on the internet to communicate with one another and route data traffic along the most appropriate path. It forms the foundation of the internet’s routing infrastructure.
What is the difference between BGP and OSPF?
OSPF is an interior routing protocol that operates within a single autonomous system and finds the shortest path. BGP is an exterior routing protocol that operates between different autonomous systems and selects the most appropriate route based on network policies rather than shortest distance.
What is BGP hijacking, and how is it prevented?
BGP hijacking occurs when a network accidentally or deliberately announces IP blocks that do not belong to it. Cryptographic validation mechanisms such as RPKI (Resource Public Key Infrastructure) and BGPsec are the primary defenses against this threat.
How many routes does the global BGP table contain?
As of 2025, the global BGP routing table contains more than 950,000 routes, a number that continues to grow year over year.
Keeping the internet’s invisible backbone operational, optimized, and secure demands genuine expertise. As a system integrator, Sekom helps carriers and enterprises not merely connect to the global internet highways — but ensure that traffic flows across them in the most efficient and secure way possible.
There is a compelling historical parallel worth noting: Türkiye established its first internet connection on April 12, 1993, from METU. Sekom had been founded exactly one year earlier, in 1992. Before Türkiye sent its first ping, Sekom had already set out with the engineering vision to build and manage that very infrastructure.
Since then, we have been designing the backbone networks of Türkiye’s major service providers and enterprises — engineering how BGP routes are drawn and how traffic flows faster and more securely through Traffic Engineering. Because in the right hands, the internet’s backbone becomes a work of engineering art.
Author: Ahmet Reçber – Network Security Operations Engineer – Sekom
Other Posts
What is BGP? Route Advertisements and the Backbone of the Internet
AI Datacenter Network Architecture | Why the Fastest GPUs Are Not Enough: The Defining Role of Network Infrastructure in AI Workloads
Meet Sekom at MWC2026 Barcelona: Network Intelligence for Real-World Operations
See all posts