Mock Exam COMP90024 2020 Semester 1

Question 1

A.) In the context of distributed databases, explain the concepts of:

Consistency [1] Availability [1]

B.) Give an example of a database technology that supports Availability in the presence of a (network) partition. [1]

C.) In the context of CouchDB clusters what is the meaning of:

Replica number [1] Number of shards [1] Read quorum [1] Write quorum [1] D.) Describe the three different Apache SPARK runtime modes:

Local [1] Cluster [1] Client [1]

Question 2

A) The NeCTAR Research Cloud is based on the OpenStack technology. Describe the role and features of the following OpenStack components: Nova [1] Horizon [1] Heat [1] Glance [1] Swift [1] Keystone [1] Neutron [1] B) Describe the interplay between these components that allows a researcher to create an instance of a virtual machine through a pre-existing snapshot from a non-public NeCTARCloud image, e.g. a snapshot created by a user. [3]

Question 3

A) Describe some of the current challenges associated with large-scale distributed systems. [4]

B) Cloud computing solves some of these issues but not all. Explain. [4]

C) What are availability zones in NeCTAR and what restrictions do they impose on NeCTARCloud-based application developers? [2]

Question 4

A) Explain Amdahl's law and discuss the challenges of its practical implementation. [2] B) The actual performance as experienced by users of shared-access HPC facilities such as the Edward cluster at the University of Melbourne can vary – where here performance can be considered as the throughput of jobs, i.e. from the time of first job submission to the time of last job completion. Explain why this can happen. [2] C) Explain how the Edward cluster has been set up to minimize this. [2] D) Explain what users can do to optimize