Cloud Computing Week 3 Assignment Answer
Cloud Computing Week 3 Assignment Answer
Course Link : Click Here
Cloud Computing Week 4
1. Which of the following is least likely to be specified in an operational cloud Service Level
Agreement(SLA)?
A. Recovery Time Objective (RTO)
B. Exact encryption algorithm and key storage procedures
C. Mean Time To Repair (MTTR) for incidents
D. Penalty / credit calculation for downtime
Answer : b
2. Which level is typically included in a multilevel cloud Service Level Agreement (SLA)?
A. Corporate/provider-wide obligations
B. Individual developer employment contracts
C. Unrelated third-party agreements
D. Hardware vendor warranties
Answer : a
3. An enterprise signs a three-tier multilevel SLA with a cloud provider. However, the corporate-level
SLA guarantees 95% uptime, the customer-level SLA specifies 98% uptime, and the service-level
SLA guarantees 99.9% uptime. In case of a service disruption, which level of SLA is the provider
LEGALLY BOUND to honor?
A. Corporate-level SLA, as it’s the broadest agreement
B. Customer-level SLA, as it’s the middle ground
C. Service-level SLA, as it’s the most specific agreement
D. The provider can choose the lowest applicable SLA
Answer : c
4. Which combination of metrics is most commonly included in production cloud SLAs to measure
availability and responsiveness?
A. Uptime percentage and mean time to respond (MTTR)
B. Developer salary and uptime percentage
C. IP address range and timezone of servers
D. Source-code commit frequency and bug count
Answer : a
5. A cloud service is guaranteed to handle 10,000 requests/hour with 99.5% success. How many failed
requests per hour are acceptable under SLA?
A. 5
B. 10
C. 50
D. 500
Answer : c
6. Which feature most clearly distinguishes parallel (analytic) databases from a single-node traditional
RDBMS in cloud analytics?
A. Support for SQL queries
B. Multi-node parallel query execution with data partitioning shared-nothing or distributed execution)
C. Transaction rollback (ACID) support
D. Use of normalized schemas only
Answer : b
7. You are designing a real-time analytics pipeline for a streaming IoT sensor network. The system must:
- Handle 1 million events per second
- Maintain sub-second latency
- Support complex windowed aggregations (sliding windows of 5 minutes)
- Which architectural choice is LEAST suitable?
A. Apache Kafka with Apache Flink for stream processing
B. Google Cloud Dataflow with shuffling disabled for specific operations
C. AWS S3 with batch processing using Apache Spark every 5 minutes
D. Azure Event Hubs with Structured Streaming
Answer : c
8. You must design a scalable pipeline to process many large video files in parallel (storage +
compute). Which storage choice is most suitable?
A. Managed relational DB (RDS/Aurora)
B. Object storage with Hadoop/Spark connector (e.g., S3/GCS + connector)
C. Single-node block volume attached to a VM
D. A small in-memory cache only
Answer : b
9. MapReduce job is designed to count word frequencies in 2 TB of text files. The mapper output is
5 TB (including overhead), and after the shuffle and sort phase, the data is 1.3 TB. During the
reduce phase, a node runs out of disk space during the merge operation. Which of the following is
the most likely cause and solution?
A. The mapper logic is inefficient; rewrite the mapper function
B. The combiner phase didn’t run; ensure combiner is explicitly configured to compress intermediate data locally
C. The shuffle and sort didn’t deduplicate keys; increase the number of reducers
D. The network bandwidth is saturated; add more nodes to the cluster
Answer : b
10. In an OpenStack deployment, if a compute node can support up to 900 virtual CPUs (vCPUs), and
the overcommit ratio for CPU is set to 4.5:1, what is the total number of CPUs that can be
allocated?
A. 4100
B. 4150
C. 4050
D. 4200
Answer : c
For other NPTEL Assignment Answer : Click Here