779 Module 6

Front 1

Column-store architecture is more efficient for data warehousing because

Back 1

Most data warehouse queries only require a few columns of data, so data stored in a column-store architecture requires less I/O.

All the data in the column has the same data type, so data can be compressed easily, making data reads more efficient.

Front 2

T/F: Row-store structure is the basic structure of RDBMS tables.

Back 2

True

Front 3

Select all that are true about MongoDB relationships.

Back 3

One way to model a one-to-many relationship in a MongoDB document is to include all the data in the same document.

MongoDB permits one-to-one and one-to-many relationships between documents.

Front 4

T/F: Eager replication is more consistent with ACIDS requirements.

Back 4

True

Front 5

Select all that are true about MongoDB.

Back 5

It would be possible to put all your data into a single MongoDB document.

MongoDB maintains some aspects of relational DBMSes, such as cursors and indexes.

It would be possible to put all your data into a single MongoDB document.

MongoDB maintains some aspects of relational DBMSes, such as cursors and indexes.

Front 6

Select all that are true of graph databases.

Back 6

Graph databases illustrate relationships between different vertices or nodes.

Graph databases are better than RDBMSes at modeling recursive relationships.

Many graph database products use both relational DBMSes and a graph database model.

Front 7

Select all that are true of key-value stores.

Back 7

Key-value stores have no schema.

Key-value stores are good choices for web applications.

Most key-value stores use eventual consistency if they implement replication.

Front 8

Select all that are true of Google's BigTable.

Back 8

A BigTable table is indexed by a row key, column key and timestamp.

Efficiency can be maximized in BigTable by grouping rows by row keys that are frequently accessed together.

Front 9

Select all that are true about BASE properties.

Back 9

“Basically available” in the context of BASE means that the system as a whole is available, even if some parts fail.

Front 10

Which of the following are correct descriptions of the differences between big data databases and relational databases?

Back 10

Big data databases do not require that data conform to a predefined schema, which is required of relational databases.

Big data database solutions are available from traditional RDBMS vendors

Front 11

Select all that are true of consistency types.

Back 11

Weak consistency may result in updates occurring in the wrong order.

Eventual consistency may result in incorrect data for some users.

Read-your-writes consistency will provide a single user with a consistent view of the data.

Front 12

Select all that are true of the CAP Theorem.

Back 12

The CAP theorem states that it is impossible to have strict consistency, availability and network partitioning simultaneously for any data model.

Front 13

Select all that are true of Hadoop.

Back 13

Hadoop works with the MapReduce paradigm, which uses key/value pairs.

Hadoop separates the data from the metadata in a master/slave architecture where the master stores the metadata and the slaves store the data.

Front 14

The reasons that ACIDS requrements do not scale well include

Back 14

ACIDS requirements include a property that ensures that all transactions are executed in the proper order. Maintaining that order becomes more complicated as the tables get larger.

ACIDS requirements include a property that ensures that all parts of a transaction are executed. This takes more time as the tables get larger and consistency must be maintained for the entire transaction.

Front 15

Some of the scalability limitations of relational databases include:

Back 15

Relational database transactions conform to ACIDS requirements which prevent them from being scalable

Relational databases with large logical tables often make use of row partitioning, using global indexes that don’t scale well

Front 16

Select all that are true of Neo4j.

Back 16

Relationships in a Neo4j graph can also include properties

student: Boston University

is an example of the properties of a graph node.

Neo4j has its own query language that is used to execute CRUD operations of Neo4j graphs.

Front 17

Which of the following are true of semi-structured data?

Back 17

The structure of semi-structured data can change rapidly.

Semi-structured data could include city and state in one tuple, and just city in another.

The structure of semi-structured data can change rapidly.

Semi-structured data could include city and state in one tuple, and just city in another

Front 18

Select all that are true about Big Data.

Back 18

Big Data databases often have no predetermined schema.

Big Data is largely comprised of structured and semi-structured data.

The hardware base for most Big Data databases is expected to fail regularly.

Front 19

Select all that are true of Dynamo DB.

Back 19

A Dynamo DB table requires a primary key.

Each attributes in a DynamoDB table is a key-value pair.

A DynamoDB table can include rows with different collections of attributes.

Front 20

Select all that are true of document-stores.

Back 20

Document-stores consist of single documents organized in collections.

It is possible to store documents in a document-store in pdf format.

Front 21

Select all that are true regarding column-store architecture.

Back 21

A column-store database may need to query a dataset by row as well as by column.

Most column-store databases support some form of the SQL language.

Front 22

Check all that are true regarding Amazon’s DynamoDB.

Back 22

A query in DynamoDB will stop if too much data is returned.

An attribute in a DynamoDB table can be multi-valued.

A DynamoDB database consists of tables with items that contain collections of attributes.

Front 23

One of the strengths of Hadoop is that data is replicated across multiple disks. Check all that are true about the process Hadoop uses for its replication.

Back 23

In Hadoop if one data node fails, the data is replicated on a different machine transparently without user intervention.

Tasks in Hadoop can be run in parallel on multiple partitions.

Front 24

Check all that are true regarding document-store architectures.

Back 24

Document-store databases may implement a flexible schema.

Document stores can support secondary indexes.

Documents in a document store are identified by keys, as is a row in a relational database.

Front 25

Check all that are true of the consistency levels used by Amazon's DynamoDB.

Back 25

DynamoDB allows the user to configure the level of consistency required.

Front 26

Which of the following are correct descriptions of why scalability is difficult for relational databases.

Back 26

As more data is added to a relational DBMS backup and recovery become difficult, because it is important to maintain the integrity of the data.

Efficient joins become more difficult as indexes grow very large.

Relational databases rely heavily on storage I/O for processing query results, causing performance degradation when the database scales much larger than available RAM.

Front 27

Performance with large databases is sometimes better on non-relational DBMS because.

Back 27

Performance with big data is sometimes better on non-relational databases because these DBMS are optimized for scalable read processing and trade consistency for improved scalability.

In big data databases data is often stored in big blocks with few if any dependencies upon other data blocks, reducing the number of I/O operations.

Most non-relational databases for big data rely on fault-tolerant frameworks implemented on low-cost commodity hardware so if one machine fails, processing can seamlessly transfer to another machine.

Front 28

Select some of the typical characteristics of “Big Data” and databases designed for big data.

Back 28

A Big Data database usually contains mostly semi-structured and unstructured data.

Processing big data relies on moving small sequences of programming code to the data and aggregating the results.

Front 29

In which of the following ways is MongoDB similar to a relational database?

Back 29

MongoDB documents contain fields which are roughly equivalent to attributes in a relational table.

MongoDB documents are part of collections in a similar way that rows are part of relational tables.

Front 30

Check all that are true of graph databases.

Back 30

Current graph databases are often combinations of graph-style links and traversal algorithms with a more traditional SQL backend. Neo4J is an exception.

Front 31

Select all that true about MapReduce.

Back 31

MapReduce can work well with column-oriented databases.

Front 32

Check all which are true regarding Hadoop.

Back 32

Hadoop is scalable. More machines can be added to the cluster to handle additional data.

The Hadoop framework is made up of the underlying file system (usually HDFS) and tools to process the data with the MapReduce paradigm.

Front 33

Check all that are true regarding eager and lazy replication in big data databases.

Back 33

Check all that are true regarding eager and lazy replication in big data databases.

If lazy replication is used, some users may read versions of the data that do not incorporate the latest updates.

Front 34

Check all that are true about the CAP theorem.

Back 34

An ACIDS-compliant relational database maintains consistency and partition tolerance, but sacrifices availability, compared to a database which does not maintain strict consistency.

A distributed relational database can maintain consistency and availability but it cannot maintain both and also support partition tolerance.

Front 35

Check all that are true about key-value stores.

Back 35

Key-value databases usually implement eventual consistency.

Front 36

Check all that are true of graph databases such as Neo4j.

Back 36

In graph databases such as Neo4j both nodes and relationships can contain key-value pairs of properties.

In graph databases such as Neo4j multiple relationships between nodes can exist simultaneously.

Neo4j uses its own query language to traverse the nodes and to perform CRUD operations.

Front 37

Select all that are true regarding Google's BigTable.

Back 37

A single cell in BigTable can contain multiple versions of the same data.

Reads and writes on single BigTable table rows are atomic.

Front 38

Check all of the following which describe ways in which MongoDB is similar to a relational database.

Back 38

MongoDB supports indexes and has its own query optimizer.

Front 39

Why is a traditional RDBMS better for transaction processing such as banking transactions than a typical NoSQL model?

Back 39

A traditional RDBMS is designed to support consistent updates and deletes, while typical big data databases are designed for scalable read performance.

Relational databases can support ACIDS transactions so that requests have the expected consistent results, which is different than the eventual consistency of most big data databases.

Relational DBMS provide better security than typical non-relational DBMS.

Front 40

An example of soft state might be

Back 40

A quiz question in this course that a student has answered in the browser but not submitted is an example of soft state.

An email draft which has not yet been saved or sent is an example of soft state.