notes.npwitknpwitk.com ↗

Quiz

Updated 4 Oct 2026

A
Quizi3: HBASE, NoSOL, VM Scheduling, Pig, Hive, Zookeeper, Name:
Nasprowich I.
ID.
6622772422
_No. 36
T/F Questions
17/18
T 1. In NWC mode, the CPUs are caps.
Sstable → Tablet y tble
T 2. Non-preemptive mode does not allow any process to interrupt the on-going process.
T 3. Credit Scheduler support automatic load balancing
F 4. Several tablets form a STable.
T 5. SS Tables store immutable data.
1, F 6. Big Table = HD
GPS
'S + Chubby + STables
F 7. There is no "join" operation in Pig Latin.
F 8. MapReduce was originally designed to adaptively support SQL syntax.
F 9. CAP theorem states that a distributed system can simultaneously guarantee Consistency, Availability, and
Partition Tolerance.
T 10. NoSQL consumes less join cost operation than relational data.
T 11. Mongo DB is a document database.
F 12. Graph database only holds the index of data on the node.
13. Pig is usually applicable for working with SQL statements.
T 14. Hive is declarative language while Pig is procedural language.
readiumte
F 15. In ZK ensemble, all request types must be forwarded to the Leader.
T 16. All servers in the ZK ensemble have the similarstate of data.
F 17. Quorum refers to the number of nodes allowed to fail and the system still keeps running.
Wth the smallest number of znode
T 18. Any members in the ensemble of Zookeeper can be promoted to be a leader
éso
63
Fill the proper word(s) into the blank(s)
14. A Scheduler decides the amount of
VCPU
15. In a credit scheduler, more credits means more _
allocated and dueetion of its allocation. (2 Points)
CPU susage
→ (time)
16. Real-time applications usually require
constraints.
Soft deadline
Hard deadline C. Round robin D. Critical
17. The three properties of the CAP theorem are consistency, eaillity, and partition taerance
3
18. Redis is primarily categorized as a key-vale store.
19. In ACID property, atomicitu none 1s.
refers to the property that either the whole process/transaction is done or
1/4

  1. GraphDB
    is NoSQL that can discover "unknown fact".
  2. Dynamo DB is a key-valre
    which is a kind NoSQL used by -
    Amazons
  3. You have real-time tasks with strict deadlines. Which scheduler is most suitable?
    A. Credit Scheduler
    C. BVT
    SEDE
    D. NWC
  4. What is the purpose of "warp" in BVT?
    (2 Points)
    A. Reduce memory usage
    C. Store data
    B. Boost priority of interactive tasks
    D. Balance load across clusters
    2 Which of the following is not true about "D Sharding"
    Each shard is independent of the DB.
    Datasets are distributed to multiple servers./
    No Join operation
    Does not require any key for sharding. v
  5. What is the main advantage of schema-less design?
    A. Enforces strict validation
    C. Slower performance
    B. Flexibility in handling diverse data
    D. Limited scalability
  6. For data management in Facebook, HDFS and
    are used as core technologies.
    A. ZooKeeper
    B. Pig
    Hbase
    D. Hive
  7. Pig Latin is considered as
    A. Declarative
    C. Object-oriented
    Procedural (data flow oriented)
    Functional only
  8. Which of the following best describes the output of a COGROUP?
    A flattened relation combining tuples from multiple datasets
    B. A relation where each key maps to grouped bags from each input relation
    C. A relation that organizes data by key into groups across multiple datasets
    D. A filtered dataset based on matching keys
    Multidimensional
  9. A data warehouse can be called as a_data lake databushich is modelled in _
    star
    consisting of a_
    fact
    table and multiple
    dimention dimension tables.
  10. In Pig, the execution only starts when the output is requested. This nature is called lazy compilation? (2Points)
  11. Which components are part of Hive architecture?
    Metastore
    Driver
    B. NameNode
    D. ZooKeeper
    E. Grunt
    schema
    2/4

2
37. Hive complies
code
into
MapReduce job and the results are then stored in
HDES
3
38. Which of the followings best describe "ZooKeeper"? And describe its role. (3 Points)
A. Resource Discovery
Service Coordinator C.Task processor D. Tasks Scheduler
39. You have a large dataset stored in HDFS and want to run SQL-like queries. Which tool is most appropriate?
A. ZooKeeper
Hive
C. Kafka D. HBase
3
40. For a 9 node cluster, we need
5
to
nodes to form a
$4
41. Which of the following are common use cases of ZooKeeper?
quorm and we can tolerate a failure of up
[2]+
eader election.
C. SQL query execution
Configuration management
D. Data warehousing
42. Which of the following statements are TRUE about ZooKeeper? (choose all that are correct)
43. ZooKeeper is designed for storing large-scale application data like HDFS.
B ZooKeeper provides strong consistency guarantees across distributed nodes.
C. ZooKeeper uses a hierarchical namespace of znodes to organize data.
. ZooKeeper is mainly used for executing MapReduce jobs.
E. ZooKeeper follows an eventual consistency model to improve performance.
ELookeeper supports leader selection protocols in distributed systems.
Short Answer:
44. Discuss the role of "Chubby" in BigTable (2 Points)
Coordinate stiff for Big Table tor example making sure that th and coordinate data to GFS
remll be one matter nade only!
45. Why do workload managers usually use PS schedulers in NWC (2 points)
-Penanto
It pronze highest throughput
even in static manager, dynamie manager
(workbad manager-
46. List TWO major advantages of graph database (2 Points)

  • Fast traverse to query the graph (very fast)
  • Having the relationship between nodes
  • Good virtvalization in Neo4j.
    3/4

Discuss TWO characteristics of Pig Latin (2 Points)

What ts the benefit of "Redo log file"? AND does it support rollback or roll-forward, explain (5 Points)

  1. Given the relations:
    Relation A (clicks):
    (apple, userl, 1)
    (apple, user2, 2)
    (banana, user3, 1)
    (banana, user4, 2)
    Relation B (ads)
    (apple, banner, 100)
    (apple, sidebar, 50)
    (banana, banner, 70) (banana, sidebar, 30)
    Write the Output of COGROUP operation. (3 Points)

41
Quiz#4: CSS451, Actor/Akka, Multi-Tenancy, Fog Computing, Cloud Security Name:
Nonprawich a、
ID
6622772422
_No.
36
T/F Questions
ไม่ได้อ่าน "
F 1. In Actor model, each actor can Send a finite number of messages to other actors
T
2. In Akka model, Actors communicate using asynchronous messages. 1
F
3. Kafka stores data only in memory. 1
F容
4. Producers consume messages from Kafka topics.
F
5. Kafka supports real-time data streaming.
TB 6. Fog computing is more decentralized than cloud computing.
1
T 7. Mist computing requires constant internet connection. x
T 8. YARN separates resource management and job execution. !
T 9. In Yarn, ResourceManager runs on every node. x T 10. In Yarn, ApplicationMaster is created per application.
F /I 11. CloudDLP enables encryption before the data is outsored to cloud.
F 12. Virtual Private Cloud is actually on-premise model but it is operated by CSP.
Fill in the blank
6. Zero-knowledge proof used in ID is genarally supports
ano nimus
identification
7. CloneCloud uses stadis analysis and dynamic profiling to decide which parts of an application to offload
8. The key advantage of cloudlets over cloud is reduced Conmunication. cost
9. Fog computing sits between Coud and mist
_ in the computing hierarchy.
10. In Yarn execution model, tasks are executed inside container managed by NodeManagers.
11. The YARN scheduler is considered a Preempt scheduler because it does not manage.
mtempt
application execution.
12. YARN supports multi tendens allowing multiple users and applications to share
13. Active bundle model consists of three major components/layers including data
metadata
_ and code or
execution run-time
14. site
chenel
aims ta exfiltrate sensitive information, including cryptographic
keys, by measuring coincidental hardware emissions.
ment
Multiple Choice
15. What is the main motivation behind using the Actor Model?
A. Improve database performance
B. Avoid concurrency issues like locks and race conditions
C. Replace distributed system D. Improve UI responsiveness E. Reduce communication cost.

  1. In the Actor Model, an actor processes messages:

A. In parallel within the same actor

B. One at a time

C. Only synchronously

D. Randomly

  1. Which of the following is NOT a characteristic of actors?

Private state

B. Shared memory

  1. Message passing

  2. Asynchronous communication E. have multiple mailboxes

  3. What does Akka primarily provide?

A. Database engine

  1. Actor-based concurrency framework
  2. Web server

D. Machine learning library E. Resource Scheduling

  1. In Akka, communication between actors is:

. Blocking

B. Synchronous

C. Asynchronous

D. File-based E. One-way only

1

  1. What is a Kafka "topic"?

A. A database table

C. A processing engine

B. A stream of records

D. A network protocol

  1. Kafka ensures scalability mainly through:

A. Encryption

Partitioning

C. Compression

Serialization

  1. In Kafka, who sends messages to topics?

A. Consumers

B. Brokers

C Producers

D. Zookeeper

E. Data sources

  1. What is the purpose of Kafka partitioning?

A. Security

B. Data compression

C. Parallelism and scalability

D. Data encryption

  1. Which message delivery guarantee may lead to duplicates?

A. At-most-once

B. Exactly-once

At-least-once

D. None

1

  1. In MCC architecture, mobile devices connect to cloud via:

  2. Bluetooth only

  3. Base stations and Internet

  4. USB cables

  5. Local storage

30.CloneCloud mainly focuses on:

A. Data storage

Offloading computation to cloud

C. Mobile UI design

D. Network routing

E. Tasks partitioning/Profiling

1

  1. Which of the following is NOT an advantage of Edge Computing?

A. Low latency

  1. Bandwidth efficiency
  2. High centralization
  • 1

Privacy enhancement

  1. Fog computing mainly helps in:

A. Increasing latency e

B. Real-time analytics

C. Eliminating cloud e

). Reducing computation

  1. What does a NodeManager in Xarn do?

A. Stores data

  1. Tracks and manages resources on a worker node
  2. Schedules jobs globally

D. Runs queries

1

  1. ApplicationMaster is responsible for:

Managing cluster storage × B. Managing a specific application lifecycle V

C. Monitoring all nodes

D. Running OS processes

E. Scheduling task

  1. Which is NOT a key function of YARN?

A. Resource allocation

B. Job scheduling

C.

Data storage

D. Multi-tenancy support

1

Short Answer

  1. A fog node processes video analytics for traffic monitoring. The system parameters are:
  • Latency requirement: ≤ 50 ms
  • Data size: High (raw video stream)
  • Bandwidth to cloud: Moderate
  • Nearby fog nodes: Available but moderately loaded

• Cloud processing time: Very fast

The system must decide whether to offload the task using Fog-to-Fog (FF) or Fog-to-Cloud

(F2C). (3 Points)

Answer: Fog-to- Fog (F2F) X

  1. What is an offset in Kafka? Why is it important? (2 Points)
  • message tracking
  • fault tor
  1. Differentiate between KStream and KTable. (3 Points)
  2. In Apache Hadoop YARN, the Resource Manager is located at the Application Master runs inside container

→ resource requests are initiated by

RAAM, and task execution is handled by

(4 Points)

What is VM overlay in cloudlet architecture? (4 Points)

VM Overlay = _

complete M

Base VM

  1. What is the main goal of ORAM?

  2. Encrypt data to prevent unauthorized access

  3. Reduce memory access latency

CHide memory access patterns from observers

D. Compress data for efficient storage

  1. In Path ORAM, where can a data block be stored?

  2. Only at the root node

  3. Only at the leaf node

  4. Anywhere along the path from root to its assigned leaf

  5. Anywhere in the entire tree

  6. What is the role of the Position Map?

  7. Store encrypted data blocks

  8. Store the physical memory address of blocks

  9. Store the logical leaf assignment of each block

  10. Store dummy blocks

  11. What type of attack does ORAM primarily defend against?

A. Brute-force attacks

Side-channel attacks (access pattern leakage)

SQL injection

D. Phishing attacks

  1. The temporary storage used during ORAM access is called

Stash

เนด้านด้วย P นะถ้าจำไม้ผิด

  1. After accessing a block, its logical position is _ leaf.

assighed x to a new random

Re Map

  1. What is the main source of overhead in ORAM? (3 Points) 3

read and write cost