Base

The target blob capacity of each block is not very high, only 380 KB, which is approximately 100 TPS in all roll ups scenarios. Using the current parameters to store blob data introduces an additional storage requirement of nearly 50GB for node operators. For a regular node, 50GB is nothing to worry about, so why not include more blobs in each block? The reason is that each consensus node still needs to download and verify all blobs. In fact, after including EIP4844, consensus clients will not consider a block to be valid until all blobs of that block are downloaded and verified. We still do not have a good data availability sampling mechanism. Therefore, it is necessary to carefully limit the additional load brought by blobs to avoid the risk of breaking the consensus mechanism by prolonging block propagation delay (note: basically, the proof period for consensus clients is 4 seconds, so each block must propagate well before 4 seconds). That's why the goal setting of blob is relatively low.
 @dfhcrd @bok