Key exchange is the problem of how two parties can establish a shared secret key that both can use to encrypt and decrypt messages, while keeping that key out of the hands of anyone listening. In symmetric encryption, the same key both locks and unlocks a message, so both parties must somehow come to hold the same key. Getting that key to both parties securely is the key distribution problem.
For most of the history of cryptography, the only answer was to deliver the key in advance through a trusted channel, such as a courier or a prearranged codebook. This worked for parties who could plan ahead but did not scale to large networks of strangers who needed to communicate on demand.
In their 1976 paper “New Directions in Cryptography,” Diffie and Hellman framed this directly, writing that the widening use of teleprocessing demanded systems that minimize the need for secure key distribution channels. Their public key distribution method, now called Diffie-Hellman key exchange, let two parties derive a shared key over an open channel using modular exponentiation, so the secret itself never crossed the wire.
Solving key exchange made secure communication between strangers practical and is the reason modern protocols such as TLS can set up an encrypted session between a browser and a server that have never interacted before.