DYOR Crypto Wiki

"The main way of identifying a block in the blockchain is via its block header hash.

The block header hash is calculated by running the block header through the SHA256 algorithm twice.

A block header hash is not sent through the network but instead is calculated by each node as part of the verification process of each block."

"Headers first is very primitive technology today. But when you join the blockchain, you download all the blockheaders which are 80 byte data objects and you try to find the longest chain. The chain part of the blockchain is entirely contained in 80 byte headers. All 600,000 of these are like less than 60 megs of data or something. So you can validate this chain, and then later you download the block data, which is much more data. So you start by downloading the headerchain, and it's low bandwidth-- you can get it over SMS or over the satellite, and then you know the right chain, and then you go find the blocks. Maybe you get it from less trustable sources, since you already have the headers nobody can lie to you about the blocks. The headers cryptographically commit to the blocks. In 2014, headers first got implemented. Before that, you had to join the network and get lucky about choosing peers because you wouldn't know that someone was lying to you until you were done with the sync."