Every BGP route carries a story—the complete history of every Autonomous System it has traversed on its journey across the Internet. This story is encoded in the AS_PATH attribute, arguably the most important attribute in all of BGP.

AS_PATH is what makes BGP a path vector protocol. It enables loop prevention, influences best path selection, provides visibility for policy decisions, and serves as a fundamental tool for traffic engineering. Understanding AS_PATH deeply is essential for anyone working with BGP, whether configuring enterprise networks, operating ISP infrastructure, or designing cloud architectures.

The AS_PATH is a well-known mandatory attribute in BGP, meaning every BGP route must include it. Its structure is more nuanced than a simple list of AS numbers.

Attribute Basics:

Path Segment Types:

AS_PATH is not just a flat list. It consists of path segments, each with a type:

Most AS_PATHs you encounter consist of a single AS_SEQUENCE segment listing the ASNs in order from most recent (leftmost) to origin (rightmost).

2-Byte vs 4-Byte AS Numbers:

Historically, AS numbers were 16 bits (0-65535). As the Internet grew, 32-bit AS numbers (0-4,294,967,295) were introduced via RFC 6793.

BGP speakers negotiate AS number size support via the 4-byte AS Number capability in OPEN messages. When a 4-byte speaker sends to a 2-byte speaker, 4-byte ASNs are replaced with the reserved value 23456 (AS_TRANS).

AS_PATH Length Calculation:

For the BGP decision process, AS_PATH length is calculated as:

• Each ASN in an AS_SEQUENCE counts as 1
• An entire AS_SET counts as 1 (regardless of how many ASNs it contains)
• Confederation segments are ignored (intra-confederation hops don't count)

This means aggregated routes (with AS_SETs) appear shorter than they actually are—a deliberate choice to encourage aggregation.

AS_PATH's most critical function is automatic loop prevention. Unlike distance vector protocols that require complex mechanisms like split horizon and poison reverse, BGP's loop prevention is elegant and foolproof.

The Loop Prevention Rule:

When a BGP router receives a route, if its own AS number appears anywhere in the AS_PATH, the route is immediately discarded.

This simple rule prevents routing loops at the AS level. If AS 65002 advertises a route to AS 65003, which then somehow advertises it back toward AS 65002, the router in AS 65002 will see its own ASN in the path and reject the route.

Loop Prevention Walkthrough:

In the diagram:

1. AS 65001 originates a route with AS_PATH = [65001]
2. AS 65002 receives it, prepends to get [65002, 65001], and advertises to AS 65003
3. AS 65003 prepends to get [65003, 65002, 65001]
4. If AS 65003 tries to advertise this back to AS 65002, the router in AS 65002 sees 65002 in the path and rejects the route

No loop is formed. The cycle is broken immediately.

Why This Works So Well:

Loop Detection vs. Loop Prevention:

It's worth distinguishing between prevention and detection:

BGP's AS_PATH provides prevention—routes that would create loops are never installed. IP's TTL mechanism provides detection—if a loop somehow forms (e.g., during convergence), TTL expiry will eventually stop the packets. Prevention is far better; TTL detection means congestion and packet loss before the problem is caught.

When a BGP router receives multiple paths to the same destination prefix, it must select the best one. AS_PATH length is a key factor in this decision, though not the most important.

The BGP Decision Process:

The complete BGP best path algorithm (simplified) considers these factors in order:

1. Highest Weight (Cisco-specific, local to router)
2. Highest Local Preference (shared within AS via iBGP)
3. Locally Originated (prefer routes this router originated)
4. Shortest AS_PATH ← AS_PATH's role
5. Lowest Origin Type (IGP < EGP < Incomplete)
6. Lowest MED (Multi-Exit Discriminator, from neighboring AS)
7. eBGP over iBGP (prefer external routes)
8. Lowest IGP metric to next-hop
9. Oldest Route (stability preference)
10. Lowest Router ID (tie-breaker)

AS_PATH length is step 4—important, but overridden by weight and local preference.

AS_PATH Length Example:

Consider a router in AS 65004 receiving these paths to 10.0.0.0/8:

Assuming equal weight and local preference, Path B wins because its AS_PATH is shortest.

Why Shorter Isn't Always Better:

AS_PATH length is a rough proxy for network distance, but it has significant limitations:

• An AS can span continents (short path, high latency)
• Links between ASes vary enormously in capacity
• Congestion isn't reflected in AS_PATH
• Business relationships aren't encoded

This is why operators use local preference (step 2) to override AS_PATH length when they have better information about path quality.

AS_PATH prepending is a technique where an AS artificially lengthens the AS_PATH to make a route less attractive. This is the primary tool for influencing inbound traffic—how traffic from the Internet reaches your network.

How Prepending Works:

When advertising a route, instead of prepending your AS once (the normal behavior), you prepend it multiple times:

The receiving AS sees a longer path and, all else being equal, will prefer alternative shorter paths.

Traffic Engineering Scenario:

An enterprise with two ISP uplinks wants most inbound traffic through ISP-A (higher capacity) with ISP-B as backup:

When Internet ASes receive both advertisements:

• Via ISP-A: AS_PATH = [65100, 65001], length = 2
• Via ISP-B: AS_PATH = [65200, 65001, 65001, 65001], length = 4

They will prefer the shorter path through ISP-A. The enterprise's inbound traffic flows primarily through ISP-A.

Prepending Limitations:

When BGP aggregates multiple prefixes into a single summary route, AS_PATH handling becomes interesting. The aggregate may combine routes from different origin ASes, so the AS_PATHs of the component routes may differ.

The Aggregation Challenge:

Consider an ISP (AS 65100) that has customers in AS 65001 and AS 65002, each announcing a /24:

• AS 65001 announces 192.0.2.0/24 with AS_PATH [65001]
• AS 65002 announces 192.0.3.0/24 with AS_PATH [65002]

The ISP wants to aggregate these into 192.0.2.0/23. What should the AS_PATH be?

AS_SET Solution:

BGP uses the AS_SET segment type to represent unordered sets of ASNs that contributed to an aggregate:

AS_SET Counting:

For best path selection, an AS_SET counts as length 1, regardless of how many ASNs it contains:

This encourages aggregation by not penalizing aggregated routes.

The ATOMIC_AGGREGATE Attribute:

When creating an aggregate, the router should set the ATOMIC_AGGREGATE attribute. This signals to downstream routers that the aggregate has lost information—specific AS_PATH details have been compressed into an AS_SET.

The AGGREGATOR attribute is also set, recording:

• The ASN of the router that created the aggregate
• The IP address of the aggregating router

These attributes help trace aggregates back to their source when troubleshooting.

AS_PATH is a powerful tool for route filtering and security. By examining the AS_PATH, routers can enforce policies about which routes to accept or reject.

AS_PATH Access Lists:

Most BGP implementations support filtering based on AS_PATH regular expressions. Common patterns:

Security Applications:

AS_PATH filtering is a frontline defense against various routing threats:

1. Prefix Leaks When an AS accidentally advertises routes it shouldn't:

• Filter by expected AS_PATH length (customers should have short paths)
• Reject paths that don't match known customer ASNs

2. Prefix Hijacks When an AS maliciously claims ownership of another's prefixes:

• RPKI validation (checks AS_PATH origin against signed ROAs)
• Filter by expected origin AS for well-known prefixes

3. Path Manipulation When an AS manipulates AS_PATH to attract traffic:

• Reject implausibly short paths (nobody has a direct path everywhere)
• Monitor for sudden AS_PATH changes on critical routes

4. Private AS Leakage When private AS numbers (64512-65534, etc.) appear in public AS_PATH:

• Filter all routes containing private ASNs
• These should never leave the originating AS

AS_PATH analysis is fundamental to diagnosing BGP routing problems. Here's how to use AS_PATH information effectively.

Reading AS_PATH Output:

Common AS_PATH Problems:

Using Looking Glasses:

BGP Looking Glass servers allow you to see routes from other networks' perspectives. When troubleshooting AS_PATH issues:

1. Find looking glasses in relevant ASes (ISPs, IXPs)
2. Query for your prefix to see the AS_PATH they receive
3. Verify that prepending/policy is having the intended effect
4. Compare paths across multiple looking glasses

Popular looking glass resources:

• Hurricane Electric (he.net)
• Route Views (routeviews.org)
• RIPE RIS (ris.ripe.net)
• Individual ISPs' public looking glasses

AS_PATH is the attribute that makes BGP a path vector protocol. Let's consolidate what we've learned:

What's Next:

In the next page, we explore BGP Attributes comprehensively. Beyond AS_PATH, BGP carries many other path attributes that influence routing decisions. You will learn about ORIGIN, NEXT_HOP, LOCAL_PREF, MED, COMMUNITY, and more—understanding how they interact in the BGP decision process.