Information Management
September 29, 2021
Time available 2:00 hours (1:30 hours for students who do not answer to Q5)
Question 1)
1. Illustrate the two phase commit protocol for the management of distributed commit.
2. Discuss the possible failures that the protocol could need to face and how they are addressed.
3. Consider the following schedule of operations performed by four processes over one variable (initially set to zero):
P1: R(X =1)R(X =2)W(X =4)
P2: W (X = 2) W (X = 3)
P3: R(X =2)R(X =1)R(X =4)R(X =3) P4: W(X = 1) R(X = 3) R(X = 4)
(a) Is the schedule sequential consistent?
If sequential consistency is satisfied, illustrate an equivalent sequence.
Otherwise, indicate the minimum number of operations (and which one) that should be removed to guarantee sequential consistency.
(b) Is the schedule causal consistent?
List all the causal dependencies in the schedule.
If the schedule it not causal consistent, indicate the operation(s) that should be removed to guarantee causal consistency.
Question 2)
1. Illustrate the ETL: Extract Transform Load process for the population of a data warehouse, briefly discussing its phases and the issues to be addressed.
2. Considering the attached table, build the bitmap index for attribute PRODUCT and the bitmap index for attribute SHOP. Write the condition operating on bitmap indexes to filter sales for products P1 and P3 sold in S1 or in S2.
Question 3)
1. Illustrate the induction strategy for building a decision tree.
2. Discuss information gain as a measure for attribute selection when building a decision tree.
3. Considering the attached table, build a decision tree using the information gain measure to choose the split attribute. For the exercise, stop at the first split and simply report the formulas to be computed for the choice.
Question 4)
With reference to NoSQL:
1. Discuss the impedance mismatch, and illustrate why NoSQL databases can represent a possible solution; 2. Discuss the BASE properties.
Question 5) only for students who did not attend database course with Prof. Samarati
1. Illustrate the adoption of two phase locking for concurrency control, clearly describing the working of two and three states lock managers.
2. Given the following schedule:
r1(y) r1(x) w1(x) r2(y) r1(x) r3(z) w3(t) w2(y) w1(t)
Tell if it could have been generated by a scheduler based on base 2PL , considering a two states lock.
Justify your answer illustrating the schedule evaluation step by step. RULES FOR THE EXAM
• Cover page: write only name, surname, matriculation number, if you did the exam with prof. Samarati
• White paper: write on each piece of paper name, surname, and matr. number (top left), number of the question (top right) • Use a different piece of paper for each question, for those questions that you decide to skip, draw a line
• Write only on one side of the paper
• Do not write close to the margins: exams will be scanned for correction
• At the end of the time or upon request, deliver your exam (the exams delivered with delay will not be accepted) • Please, order the pieces of paper you are delivering for the exam as follows:
– cover page
– answers (in increasing order of question)
Matr: Question
Surname, Name:
DB with prof. Samarati?: (Y) (N)
Final Mark:
Information Management – September 29, 2021
Matr: Surname, Name: Question 2.2
1 2 3 4 5 6 7 8
Surname, Name:
Question 3.3
1 manager 1 y
2 manager 1 y
3 manager 2 y
4 employee 1 n
5 employee 2 n
6 director 3 y
7 manager 2 n
8 director 3 n
9 employee 3 y
10 manager 1 y