A router is a networking device that forwards data packets between computer networks. Routers perform the traffic directing functions on the Internet. Data sent through the internet, such as a web page or email, is in the form of data packets. A packet is typically forwarded from one router to another router through the networks that constitute an internetwork (e.g. the Internet) until it reaches its destination node.

A router is connected to two or more data lines from different IP networks. When a data packet comes in on one of the lines, the router reads the network address information in the packet header to determine the ultimate destination. Then, using information in its routing table or routing policy, it directs the packet to the next network on its journey.

The most familiar type of IP routers are home and small office routers that simply forward IP packets between the home computers and the Internet. An example of a router would be the owner's cable or DSL router, which connects to the Internet through an Internet service provider (ISP). More sophisticated routers, such as enterprise routers, connect large business or ISP networks up to the powerful core routers that forward data at high speed along the optical fiber lines of the Internet backbone.

The main purpose of a router is to connect multiple networks and forward packets destined either for its own networks or other networks. A router is considered a layer-3 device because its primary forwarding decision is based on the information in the layer-3 IP packet, specifically the destination IP address. When a router receives a packet, it searches its routing table to find the best match between the destination IP address of the packet and one of the addresses in the routing table. Once a match is found, the packet is encapsulated in the layer-2 data link frame for the outgoing interface indicated in the table entry. A router typically does not look into the packet payload,[citation needed] but only at the layer-3 addresses to make a forwarding decision, plus optionally other information in the header for hints on, for example, quality of service (QoS). For pure IP forwarding, a router is designed to minimize the state information associated with individual packets. Once a packet is forwarded, the router does not retain any historical information about the packet.

The routing table itself can contain information derived from a variety of sources, such as a default or static routes that are configured manually, or dynamic routing protocols where the router learns routes from other routers. A default route is one that is used to route all traffic whose destination does not otherwise appear in the routing table; this is common – even necessary – in small networks, such as a home or small business where the default route simply sends all non-local traffic to the Internet service provider. The default route can be manually configured (as a static route), or learned by dynamic routing protocols, or be obtained by DHCP.

A router can run more than one routing protocol at a time, particularly if it serves as an autonomous system border router between parts of a network that run different routing protocols; if it does so, then redistribution may be used (usually selectively) to share information between the different protocols running on the same router.

Besides making a decision as to which interface a packet is forwarded to, which is handled primarily via the routing table, a router also has to manage congestion when packets arrive at a rate higher than the router can process. Three policies commonly used in the Internet are tail drop, random early detection (RED), and weighted random early detection (WRED). Tail drop is the simplest and most easily implemented; the router simply drops new incoming packets once the length of the queue exceeds the size of the buffers in the router. RED probabilistically drops datagrams early when the queue exceeds a pre-configured portion of the buffer, until a pre-determined max, when it becomes tail drop. WRED requires a weight on the average queue size to act upon when the traffic is about to exceed the pre-configured size, so that short bursts will not trigger random drops.

Another function a router performs is to decide which packet should be processed first when multiple queues exist. This is managed through QoS, which is critical when Voice over IP is deployed, so as not to introduce excessive latency.

Yet another function a router performs is called policy-based routing where special rules are constructed to override the rules derived from the routing table when a packet forwarding decision is made.

Router functions may be performed through the same internal paths that the packets travel inside the router. Some of the functions may be performed through an application-specific integrated circuit (ASIC) to avoid overhead of scheduling CPU time to process the packets. Others may have to be performed through the CPU as these packets need special attention that cannot be handled by an ASIC.