Common vs UE-specific
Common Search Space (CSS) carries system-wide DCI — SIB scheduling, RAR, paging, slot-format and power-control — addressed by group RNTIs, located at fixed candidates everyone can find. UE-specific Search Space (USS) carries your unicast grants (C-RNTI), with candidate locations hashed from your RNTI so UEs spread out.
Common (CSS)
UE-specific (USS)
Carries C-RNTI / CS-RNTI / MCS-C-RNTI DCI for your PDSCH & PUSCH grants. Candidate CCE positions are derived from a per-UE hash, so two UEs in the same CORESET rarely fully collide.
Yp,n = (Ap·Yp,n−1) mod 65537, Yp,−1 = nRNTI
Candidates across aggregation levels
Each search space set lists how many candidates to monitor per aggregation level. For USS, the CCE start of each candidate comes from the hash above — change the RNTI and watch the candidates relocate within the CCE pool.
Monitoring occasions
A search space set monitors PDCCH on a periodicity of ks slots, starting at an offset, for a duration of consecutive slots. Outside those occasions the UE's PDCCH receiver can sleep — that's the basis of power saving.
Blind-decode & CCE budget
Per slot the UE can only monitor so many candidates and estimate so many CCEs — limits set by sub-carrier spacing (TS 38.213 Tables 10.1-2/3). Pile on candidates and watch the budget; over-budget search spaces are dropped by rule.