Understanding TCP Connection Limits and the Role of Port Numbers

In TCP communication a connection is uniquely identified by a four‑tuple {local IP, local port, remote IP, remote port}. The port number occupies 16 bits, giving 2^16 (65536) possible values, with port 0 reserved, so each host can use up to 65535 distinct ports.

When a machine acts only as a client, it can open at most 65535 simultaneous connections because each outbound connection requires an exclusive local port. When the machine acts as a server, the listening port is fixed, and the remote IP and remote port vary, allowing a theoretical maximum of 2^32 (IP addresses) × 2^16 (ports) ≈ 2^48 concurrent connections.

In practice the actual number of concurrent TCP connections is limited by system resources such as memory and the maximum number of open file descriptors, not by the port count. Linux servers can often support over 100 000 or even millions of connections after tuning memory and descriptor limits.

Port reuse is possible: multiple clients can connect to the same server port (e.g., HTTP port 80) because each connection is distinguished by the client’s IP and port. TCP and UDP sockets can share the same port number because they use different protocol identifiers, forming a five‑tuple for full uniqueness.

The article also outlines the typical TCP handshake process (listen, connect, accept, communicate) and emphasizes that large‑scale services rely on server clusters, load balancing, and resource monitoring to handle billions of requests without exhausting a single machine’s resources.