Reachability from Different Locations

If you've ever traveled internationally and found that a website works perfectly from home but fails to load abroad—or heard from a colleague in another country who can access a service you cannot—you've discovered something fundamental about the Internet: it isn't one network. It's thousands of independent networks, loosely stitched together, each making its own decisions about how to handle your traffic.

Your packets don't travel "the Internet." They travel through a chain of strangers, each deciding independently whether to help you reach your destination.

Why Location Matters

There's no single "correct" path between two points on the Internet. Each network along the way makes independent decisions about where to forward traffic, based on:

Routing policies: Different ISPs have different agreements about exchanging traffic. An ISP in Europe might route your packets through London; an ISP in Asia might route through Singapore.

Network topology: The physical infrastructure connecting networks differs by region. Some areas have abundant connectivity; others depend on a handful of critical links.

Administrative decisions: Countries, organizations, and networks implement filters and policies that affect what traffic they allow.

Geographic optimization: Content delivery networks and cloud providers direct traffic to nearby servers, so users in different locations literally connect to different machines.

Routing Path Failures

A website might be unreachable from one location while working perfectly from another—even though the destination server is healthy. The problem isn't the server; it's the path.

Imagine your ISP in New York routes traffic to a European server through an undersea cable that's been damaged. Meanwhile, users in California connect through a different cable and experience no issues. Same destination, different paths, different outcomes.

This happens because Internet routing isn't centrally coordinated. Each autonomous system makes local decisions. When one path breaks, it doesn't automatically affect all paths.

Peering and Transit

ISPs establish peering relationships—agreements to exchange traffic directly—with some networks but not others. When two ISPs don't peer directly, traffic between them must transit through intermediaries.

This creates situations where users of ISP A reach a service instantly via direct peering, users of ISP B route through multiple transit providers (adding latency and failure points), and users of ISP C find the service completely unreachable because no transit path exists.

The service hasn't changed. The networks between you and the service determine your experience.

Geographic Content Restrictions

Some reachability differences are intentional. Services restrict access based on location for business or legal reasons:

Streaming services enforce licensing restrictions by blocking certain countries. Financial services comply with regulations by limiting access to specific regions. Cloud providers disable services where they lack proper licensing.

These restrictions typically use geolocation databases that map IP addresses to countries, then block or redirect traffic accordingly. From the user's perspective, the service simply doesn't work—but it's working exactly as designed.

Censorship and Filtering

Governments and organizations implement filtering that varies by location:

National firewalls block specific domains, IP addresses, or protocols at the country level. Corporate networks implement different policies at different offices. ISPs filter content based on local regulations.

When filtering occurs at the network level, users within that network see destinations as unreachable, while users outside the filtered network access the same resources without issues. The destination hasn't moved. The network between you and the destination has decided you shouldn't reach it.

BGP: The Internet's Unreliable Postal System

Border Gateway Protocol (BGP) is how networks tell each other what destinations they can reach. But BGP announcements don't propagate uniformly across the Internet.

A destination network might announce its IP prefixes to some peers but not others. BGP communities can tag routes for special handling, causing different ISPs to treat the same destination differently. Route filtering by ISPs means valid BGP announcements might be ignored in parts of the Internet.

The result: some networks know how to reach a destination while others genuinely don't know it exists.

DNS: The First Point of Failure

Reachability issues often start before your packets try to reach their destination. DNS resolution—converting domain names to IP addresses—varies by location.

Geo-DNS returns different IP addresses based on where you're querying from. Users in North America receive IPs for North American servers; users in Europe get European servers. If one region's servers have problems, only users directed to that region are affected.

DNS filtering blocks access to specific domains. ISP-level filtering redirects blocked domains to error pages. Enterprise filtering blocks categories of content. National filtering implements government censorship. When this happens, name resolution fails—but only for users on that particular network.

DNS propagation delays mean that when records change, updates don't reach all DNS servers simultaneously. Some users see new IP addresses immediately; others continue resolving to old addresses that may no longer work.

Diagnosing Location-Based Issues

When something works from one location but not another, compare the experience from both sides:

Compare routes using traceroute from both locations. Different paths or failures at specific hops indicate routing problems.

Check DNS responses by querying from both locations. Different IP addresses suggest geo-DNS or different infrastructure.

Test specific IPs by connecting directly to IP addresses instead of domain names. If this works when the domain doesn't, DNS filtering or misconfiguration is the culprit.

Examine BGP visibility using looking glass services. If one location sees a BGP route and another doesn't, the destination isn't announcing routes to part of the Internet.

Real-World Examples

When a cloud region experiences an outage, only users routed to that region are affected. AWS US-East-1 goes down; geo-DNS continues directing US users to the failed region; European users routed to EU-West-1 experience nothing unusual.

When BGP hijacking occurs—malicious or accidental announcements of someone else's routes—destinations become unreachable from specific parts of the Internet. ISPs closer to the hijacker accept the false routes; distant ISPs continue using legitimate ones.

When CDN nodes become overloaded, users near those nodes experience failures while users routed to healthier nodes continue unaffected.

Improving Reachability

Multihoming: Announcing your network through multiple upstream ISPs creates diverse routing paths, reducing dependence on any single provider.

Anycast: The same IP address announced from multiple locations. If one location becomes unreachable, traffic automatically routes to the next-nearest.

Geographic monitoring: Check reachability from diverse locations. A service might appear healthy from your monitoring location while being unreachable elsewhere.

Intelligent DNS: Detect failures and stop directing traffic to unhealthy locations, automatically routing users to working infrastructure.

The Internet has no center, no authority, no guarantee that any two points can reach each other. When you can't access something from where you are, it's not necessarily broken—it's that the particular chain of networks between you and your destination has a gap. Someone else's chain might be fine.

Understanding this helps you diagnose problems and design services that remain reachable even when parts of the Internet fail.