A computational backlog is an accumulation of processing work, such as calculations, transactions or messages, that a computer system has received but not yet completed. It builds up whenever work arrives faster than the system can handle it, for example during a sudden spike in demand, after an outage or when hardware is too slow for the load. Work typically waits in a message queue until capacity frees up.
In blockchain networks the idea appears in two places. A node that has been offline must download and verify all the blocks it missed before it is in sync, and a new full node faces the largest backlog of all, the entire chain history. At the network level, when more transactions are submitted than blocks can hold, unconfirmed transactions pile up in the mempool, a form of network congestion. Backlogs are cleared by adding capacity, spreading work across machines or waiting for demand to fall, at a pace set by the system's throughput.
Ref 1300000286