Telephony Routing vs Internet Routing
Telephony Routing
Circuit-Switched Nature
- Telephony routing is based on circuit-switched networks, where a dedicated communication path (circuit) is established between two endpoints before any data (voice) transmission occurs.
- The circuit is reserved for the duration of the call, guaranteeing a consistent and dedicated bandwidth.
- Once the call ends, the circuit is released and can be used by other communications.
Hierarchical Structure
- Telephony networks use a 3-level hierarchy:
- Local Exchange Carriers (LECs) manage local calls and direct traffic within the same geographic area.
- Toll switches handle long-distance calls, determining whether to send the call over a one-hop or two-hop path between toll switches.
- The core network is fully connected, consisting of central switches that route calls across long distances.
Routing Algorithm
- Calls are routed based on proximity and resource availability:
- If the call is within the same Central Office (CO), it is directly connected.
- If it’s within the same Local Exchange Carrier (LEC) but across different COs, a one-hop path is used.
- If it’s an external call, the toll switch determines the optimal route. If the direct route is unavailable, a two-hop path is selected.
Resource Reservation
- Resource reservation is fundamental to telephony routing. This includes:
- Reserving a 64 kbps channel for each voice connection.
- Signaling protocols are used to set up, maintain, and tear down the connection.
- The path and the required resources (bandwidth) are reserved simultaneously, ensuring no packet loss or jitter.
Stable Network Topology
- The telephony network is designed for voice traffic, which is highly predictable and stable throughout the day.
- Load balancing is planned in advance based on expected call volumes at different times.
- Network routes and resources are pre-planned, and routes can be chosen optimally in advance because the network topology changes very little.
Technology and Economic Aspects
- Extremely reliable switches ensure minimal downtime, typically less than a few minutes per year.
- The system was built with dumb end-systems (phones) because computing power was limited in early telephony networks, so all intelligence was placed in the network.
- The network’s reliability is driven by its components’ reliability, which ensures continuous service.
Centralized Control
- The telephony network is often controlled by a single organization (e.g., AT&T for the US).
- This control allows global optimization of routing and network usage, as the organization can collect statistics and implement changes across the entire network.
Internet Routing
Packet-Switched Nature
- Internet routing is based on packet-switched networks, where data is broken into smaller packets that are transmitted independently across the network.
- Unlike telephony, no dedicated path is established before communication; instead, each packet may follow a different route to the destination.
- Packet-switched networks are designed for data communication, where paths can change dynamically based on network conditions.
Decentralized and Dynamic Structure
- The Internet is built on a decentralized model, meaning there is no single organization that controls the entire network.
- It consists of many autonomous systems (AS) that interconnect with each other through peering agreements.
- Routing decisions are made using distributed algorithms (e.g., BGP, OSPF), and the network topology can change frequently.
Routing Algorithm
- Routing on the Internet uses dynamic distributed protocols to compute the best paths based on factors such as hop count, link costs, and congestion.
- Internet routing is policy-driven, especially between autonomous systems. This means that routes may not always be the shortest but are determined by business relationships and agreements between networks.
No Resource Reservation
- The Internet does not reserve resources like telephony. Instead, packets are sent whenever the network has capacity.
- This means that packet loss, jitter, and delays can occur during periods of high traffic.
- While Quality of Service (QoS) techniques can be used to prioritize certain types of traffic (e.g., voice or video), there is no guaranteed bandwidth as in telephony routing.
Unreliable and Changing Topology
- Internet components (e.g., routers, links) are cheap and unreliable compared to telephony networks.
- As a result, the topology changes frequently due to link failures, congestion, or other issues.
- The Internet needs dynamic routing to continuously adapt to these changes, ensuring that packets can still find a route to their destination.
Technology and Economic Aspects
- The Internet was originally an overlay on top of the telephone infrastructure and has since evolved into a global communication network.
- The use of cheap and unreliable components made it affordable for many entities to interconnect their local networks and build the global Internet.
- Because the Internet is built on top of multiple administrative domains (autonomous systems), each domain manages its own internal routing while participating in global routing.
Decentralized Control
- The Internet operates with multiple independent organizations, each controlling its part of the network.
- This leads to multiple routing policies and the need for protocols like BGP to manage routing between different organizations.
- Internet routing is less optimized than telephony because global optimization is challenging when control is spread across different entities.
Key Differences
| Feature | Telephony Routing | Internet Routing |
|---|---|---|
| Network Type | Circuit-switched | Packet-switched |
| Connection Setup | Requires connection setup before transmission | No connection setup; packets routed individually |
| Routing Algorithm | Static or semi-static, hierarchical | Dynamic, distributed algorithms |
| Resource Reservation | Resources (e.g., bandwidth) are reserved | No resource reservation, best-effort transmission |
| Reliability | Highly reliable, minimal downtime | Less reliable, frequent topology changes |
| Topology | Stable, predictable traffic | Dynamic, frequent changes |
| Control | Centralized (single organization) | Decentralized (many independent networks) |
| Policy | Less flexibility; focus on efficiency | Policy-driven, influenced by business relationships |
| Use Case | Voice communication | Data communication, multimedia, web services |
Telephony routing is highly centralized, reliable, and predictable, designed for stable voice traffic, while Internet routing is decentralized, dynamic, and flexible, designed to handle the unpredictability of modern data communication across a global, heterogeneous network.