Learning objectives
What you will be able to explain
- Which two locality types are highlighted as central for caching effectiveness?
- Which formula for cache-hit-ratio is shown in the storage lecture?
- What does the virtual-memory slide explicitly state?
- How are pages characterized in the data-layout lecture?
- In slotted pages, how is tupleID represented?
- What is a direct consequence of DSM/column layout mentioned in the slides?
Section 01
Which two locality types are highlighted as central for caching effectiveness? to What does the virtual-memory slide explicitly state?
01
Which two locality types are highlighted as central for caching effectiveness?
The slide introduces temporal and spatial locality as the key access localities.
Source: 01_Storage.pdf, page 10
02
Which formula for cache-hit-ratio is shown in the storage lecture?
The cache-hit-ratio is defined as nb_hits divided by total accesses (hits plus misses).
Source: 01_Storage.pdf, page 10
03
What does the virtual-memory slide explicitly state?
The lecture states virtual memory can exceed physical memory and may page to disk.
Source: 01_Storage.pdf, page 20
Section 02
How are pages characterized in the data-layout lecture? to What is a direct consequence of DSM/column layout mentioned in the slides?
01
How are pages characterized in the data-layout lecture?
Pages are described as fixed-sized blocks with a header and data portion.
Source: 02_Data_Layouts.pdf, page 10
02
In slotted pages, how is tupleID represented?
The slotted-page structure explicitly shows tupleID = pageID + slotID.
Source: 02_Data_Layouts.pdf, page 11
03
What is a direct consequence of DSM/column layout mentioned in the slides?
The column-layout slide states that column layout requires tuple reconstruction.
Source: 02_Data_Layouts.pdf, page 31
Section 03
Which NSM/DSM/PAX comparison is consistent with the table in the lecture? to How is CLOCK (Second Chance) characterized?
01
Which NSM/DSM/PAX comparison is consistent with the table in the lecture?
The table lists inter-record spatial locality as No for NSM, Yes for DSM, Yes for PAX.
Source: 02_Data_Layouts.pdf, page 45
02
Which policy is based on recency according to the eviction-policy slide?
LRU is described as eviction based on time/recency.
Source: 03_Caching.pdf, pages 20 and 22
03
How is CLOCK (Second Chance) characterized?
The lecture explicitly says CLOCK improves over FIFO and approximates LRU with less overhead.
Source: 03_Caching.pdf, page 25
Section 04
During CLOCK eviction search, what happens when a page with bit=1 is encountered? to In indexing basics, when is an index usually most useful?
01
During CLOCK eviction search, what happens when a page with bit=1 is encountered?
The algorithm states: if bit is 1, set it to 0 and continue; bit 0 is evicted.
Source: 03_Caching.pdf, page 26
02
What is the key idea behind prefetching pages in the caching lecture?
Prefetching is described as loading data not needed yet but likely soon, often using sequential/asynchronous reads.
Source: 03_Caching.pdf, page 33
03
In indexing basics, when is an index usually most useful?
The selectivity discussion shows index vs scan depends on expected result size; indexes are better for few qualifying tuples.
Source: 04_Indexing 1.pdf, page 8
Section 05
What is a dense index? to What is true for a secondary index in this lecture?
01
What is a dense index?
Dense index is defined as one index entry per record/tuple.
Source: 04_Indexing 1.pdf, pages 12 and 14
02
Which statement about sparse indexes is correct?
Sparse index is shown as a strict subset (often one per block) and useful when records are sorted.
Source: 04_Indexing 1.pdf, pages 12 and 16
03
What is true for a secondary index in this lecture?
Secondary index is described as not determining data order; the slide states it can only be dense.
Source: 04_Indexing 1.pdf, page 23
Section 06
Which balancing property is stated for B-trees? to What can happen during B-tree insertion if there is no room?
01
Which balancing property is stated for B-trees?
The lecture states B-trees are always balanced with equal root-to-leaf path length.
Source: 05_Indexing 2.pdf, page 12
02
What design goal is emphasized for B-trees in DBMS?
B-trees are presented as designed for block-wise access; nodes are kept between half-full and full.
Source: 05_Indexing 2.pdf, page 12
03
What can happen during B-tree insertion if there is no room?
Insertion without room triggers split; recursive propagation can reach root and increase tree height.
Source: 05_Indexing 2.pdf, page 29
Section 07
Which deletion behavior is shown for B+-trees? to Which two collision-handling strategies are explicitly listed?
01
Which deletion behavior is shown for B+-trees?
The delete slide describes stealing keys from siblings or merging if stealing is not possible, with possible propagation.
Source: 05_Indexing 2.pdf, page 35
02
What is the range of hash function output h(K) in the lecture’s basic model?
Hash output is defined as an integer between 0 and B-1, with B buckets.
Source: 06_Indexing 3_hash_tables.pdf, page 8
03
Which two collision-handling strategies are explicitly listed?
The collision slide names separate chaining and open addressing as handling strategies.
Source: 06_Indexing 3_hash_tables.pdf, page 9
Section 08
Why is dynamic resizing discussed for hash tables? to Which set is shown as kinds of multidimensional indexes?
01
Why is dynamic resizing discussed for hash tables?
The lecture motivates resizing due to fill/waste tradeoffs and notes full rehashing is costly.
Source: 06_Indexing 3_hash_tables.pdf, page 11
02
Which dynamic hash-table families are covered in the lecture?
The static-vs-dynamic slide explicitly lists extensible hashing and linear hashing.
Source: 06_Indexing 3_hash_tables.pdf, page 20
03
Which set is shown as kinds of multidimensional indexes?
The slide’s index-family overview lists these multidimensional structures.
Source: 07_1_Multidimensional Indexing.pdf, page 63
Section 09
How are bitmap indexes described conceptually? to How is point search explained for the grid file?
01
How are bitmap indexes described conceptually?
The bitmap-index slide defines a bit vector per value showing whether it is present in each row.
Source: 07_1_Multidimensional Indexing.pdf, page 92
02
What is the core construction idea of a grid file?
The grid-file idea slide states exactly this stripe-and-grid construction.
Source: 07_2_multidimensional indexing_pt2.pdf, page 2
03
How is point search explained for the grid file?
The search slide describes determining per-dimension positions and mapping to a bucket, with overflow blocks permitted.
Source: 07_2_multidimensional indexing_pt2.pdf, page 3
Section 10
Which statement matches the kd-tree slide? to What is emphasized about logical vs. physical query operators?
01
Which statement matches the kd-tree slide?
The kd-tree slide explicitly defines them as a BST generalization over k dimensions with attribute alternation.
Source: 07_2_multidimensional indexing_pt2.pdf, page 20
02
For an R-tree point query, what is required when regions overlap?
The operations slide states that if a point is contained in multiple regions, descent must continue through all overlaps.
Source: 07_2_multidimensional indexing_pt2.pdf, page 40
03
What is emphasized about logical vs. physical query operators?
The lecture gives hash join and sort-merge join as alternative physical implementations.
Source: 08_Query Execution.pdf, page 5
Section 11
Which dimensions of processing models are listed in the query-execution lecture? to What is the parser’s responsibility in query optimization?
01
Which dimensions of processing models are listed in the query-execution lecture?
These three dimensions are explicitly listed on the processing-model slide.
Source: 08_Query Execution.pdf, page 10
02
How is morsel-driven parallelism described?
The slide defines morsels (around 1000 tuples) and dynamic runtime dispatch to threads.
Source: 08_Query Execution.pdf, page 62
03
What is the parser’s responsibility in query optimization?
Parser responsibilities shown are parse-tree construction and syntax checking.
Source: 09_Query_Optimization.pdf, page 8
Section 12
Which is a preprocessing task (not parser-only)? to What assumption is called out as strong when combining conjunctive selectivities?
01
Which is a preprocessing task (not parser-only)?
Preprocessing includes semantic correctness checks like relation, attribute, and type validation.
Source: 09_Query_Optimization.pdf, page 11
02
How is output cardinality linked to selectivity in the lecture?
The slide states output cardinality equals input cardinality multiplied by selectivity (filter factor).
Source: 09_Query_Optimization.pdf, page 26
03
What assumption is called out as strong when combining conjunctive selectivities?
The conjunction slide notes multiplication of selectivities relies on an independence assumption and calls it strong.
Source: 09_Query_Optimization.pdf, page 29
Section 13
What is the Selinger-style enhancement over plain DP in the slides? to When is a schedule conflict-serializable?
01
What is the Selinger-style enhancement over plain DP in the slides?
Selinger-style optimization keeps track of interesting orders beyond the single cheapest partial plan.
Source: 09_Query_Optimization.pdf, page 49
02
What does 'serializable' mean in the concurrency lecture?
Serializable schedules are defined as producing the same result as a serial schedule; interleaving can still occur.
Source: 10_Concurrency Control.pdf, page 24
03
When is a schedule conflict-serializable?
The lecture states conflict-serializable iff the precedence graph has no cycle.
Source: 10_Concurrency Control.pdf, page 36
Section 14
What does 2PL guarantee, and what can still occur? to What is true for READ COMMITTED according to the isolation-level slide?
01
What does 2PL guarantee, and what can still occur?
2PL is presented as guaranteeing serializability, while deadlocks remain possible.
Source: 10_Concurrency Control.pdf, page 47
02
Which statement correctly describes Strict 2PL in the slides?
Strict 2PL keeps exclusive locks until commit/abort acknowledgement and ensures recoverable schedules.
Source: 10_Concurrency Control.pdf, page 53
03
What is true for READ COMMITTED according to the isolation-level slide?
READ COMMITTED disallows dirty reads but still allows fuzzy reads and phantoms.
Source: 10_Concurrency Control.pdf, page 70
Section 15
What does the STEAL policy permit? to Which WAL rule is stated for update logging?
01
What does the STEAL policy permit?
STEAL allows writing dirty pages even for uncommitted transactions.
Source: 11_Recovery.pdf, page 7
02
Which method/policy pairing is shown in the recovery overview?
The overview slide pairs shadow paging with FORCE + NO-STEAL and WAL with NO-FORCE + STEAL.
Source: 11_Recovery.pdf, page 10
03
Which WAL rule is stated for update logging?
WAL requires forcing update log records before writing the affected data page; commit also requires log durability.
Source: 11_Recovery.pdf, page 16
Section 16
What is the three-pass recovery order shown in the lecture? to What characterizes a homogeneous distributed database in the lecture?
01
What is the three-pass recovery order shown in the lecture?
The recovery process is shown as Analysis pass, then Redo pass, then Undo pass.
Source: 11_Recovery.pdf, page 24
02
How is speedup defined in the parallel-databases lecture?
Speedup is defined as sequential runtime over parallel runtime on p nodes; linear speedup is the ideal.
Source: 12_Parallel_DBs_and_Big_data.pdf, page 7
03
What characterizes a homogeneous distributed database in the lecture?
Homogeneous systems are described with identical software and cooperative behavior, appearing as one system to users.
Source: 12_Parallel_DBs_and_Big_data.pdf, page 15
Section 17
Which mapping between parallelism mode and workload focus is stated? to How are replication and fragmentation described for distributed storage?
01
Which mapping between parallelism mode and workload focus is stated?
The slide frames inter-query parallelism as key for transactional scenarios and intra-query parallelism for analytical scenarios.
Source: 12_Parallel_DBs_and_Big_data.pdf, page 24
02
How are replication and fragmentation described for distributed storage?
The lecture defines replication vs fragmentation and explicitly notes that both can be combined.
Source: 12_Parallel_DBs_and_Big_data.pdf, page 31
Knowledge check
Turn understanding into recall.
The quiz now follows the same concepts in scored form. You can return to this lesson from the quiz whenever a gap appears.