Lecture 12 & 13

Question 1

What is NoSQL Database?

Answer 1

Databases that do not follow the traditional Relational Model.

Question 2

Why do we need NoSQL Databases?

Answer 2

To help with Big Data, in order to help with Horizontal and Vertical scaling of the Database

Question 3

What is Horizontal scaling?

Answer 3

Adding more nodes to a system

Question 4

What is Vertical scaling?

Answer 4

Adding more resources to a node

Question 5

What is Availability?

Answer 5

a database always responding to queries

Question 6

What is Consistency?

Answer 6

a database gives the same response at queries happening at the same time

Question 7

what is Distributed Databases?

Answer 7

Splitting computational Load among different nodes

Question 8

What is a Cluster?

Answer 8

The set of computers that co-operate to manage the Database

Question 9

What is Two-phase Commit?

Answer 9

The algorithm used to enforce consistency during transactions

Question 10

What is a Transaction?

Answer 10

a set of changes in a database that is treated as a single change

Question 11

What does MVCC stand for?

Answer 11

Multi-Version Concurrency Control

Question 12

What is MVCC?

Answer 12

it is a method that stores data in various versions to ensure availability and recovery of data from a partition by reconciling the single databases with revisions (data isn’t replaced, just given a new revision number)

Question 13

What is Partition/Node Failure?

Answer 13

When a node in a cluster fails causing it’s data to become outdated.

Question 14

What is MVCC used for?

Answer 14

Coarse-grained DBMS models like document-oriented DBMS e.g. CouchDB

Question 15

What is Sharding?

Answer 15

The Horizontal partitioning of a databse i.e. the rows are partitioned in a subset that is stored on different servers.

Question 16

Why do we use Sharding?

Answer 16

Allows better performance by distributed computer loads.

Question 17

What kinds of Sharding are there?

Answer 17

Hash Sharding and Range Sharding

Question 18

What is Hash sharding?

Answer 18

distributing rows evenly across the cluster

Question 19

What is Range Sharding?

Answer 19

Similar rows (e.g. tweets from same area) stored on the same node

Question 20

What is Replication?

Answer 20

The action of storing the same row on different node to have fault-tolerence

Question 21

What are MapReduce Algorithms?

Answer 21

A paradigm suited to parallel computing of the Single-Instruction, Multiple-Data type.

Question 22

How does MapReduce Work?

Answer 22

1. Map: distribute data across machines

2. Reduce: hierarchically summaries data until a result is obtained.

Question 23

Example of MapReduce?

Answer 23

See Billy Boy Diagram in Lecture 12

Question 24

What is a document database? (e.g. CouchDB)

Answer 24

A Type of Database where data is stored in documents expressed as JSON

Question 25

How is data queried, added, modified and deleted?

Answer 25

HTTP requests: the protocol every web-browser uses to download webpages from a server GET method: retrieve data from server POST method: Create content on server PUT method: Update/create data on server when ID is known DELETE method: remove data from server.

Question 26

What is id of a document?

Answer 26

is a unique identifier that can be specified, by default is generated by CouchDB and is guaranteed to be unique.

Question 27

What is revision number?

Answer 27

An identifier to differ the different versions of a document. every database instance will pick up the same revision as the “live” version of the document

Question 28

What does UUID stand for?

Answer 28

Universal Unique IDentifier

Question 29

What is UUID?

Answer 29

is a sequence of letters and number that are guaranteed to be unique

Question 30

What happens when conflicts occur in CouchDB?

Answer 30

When a document recieves different updates then two different revisions are added. however only one revision is returned. The "wining" revision is guaranteed to be the same on any node of the cluster.

Question 31

Why is replication useful?

Answer 31

Can be combined with sharding with the objective of maximizing availability while maintaining a minimum level of data safety.

Question 32

Why is Map Reduce useful?

Answer 32

As it is horizontally scalable, MapReduce is the tool of choice when operations on big datasets are to be done