02 Distributed Storage Flashcards
1
Q
What is fragmentation?
A
Splitting data intro fragments to distribute them to different nodes
2
Q
What is replication?
A
Put same data fragment on multiple nodes
3
Q
Horizontal Fragmenation
A
Split data by rows
4
Q
Vertical fragmentation
A
Split data by columns
5
Q
Shared Data Advantages (2)
A
- Load Balancing
- Fault-tolerant
6
Q
Shared Data Disadvantages
A
High Latency
7
Q
Shared-nothing advantages (2)
A
- Low latency
- simple to implement
8
Q
Shared-nothing disadvantage
A
Prone to load imbalance and node failures
9
Q
What is allocation of data?
A
Decide on which server to put fragment
10
Q
3 Fragmentation Schemes
A
- Round-Robin
- Range partitioning
- Hash-Partitioning
11
Q
What is co-partitioned data?
A
Same partitioning function used for both tables
12
Q
What is pruning of partitions?
A
- skip unnecessary partitions during query execution
13
Q
What is data-distribution skew?
A
Some nodes have more data than others
14
Q
A