07 02 11

Page 1

Data Bases 2 Prof. S. Paraboschi – February 7th, 2011 – Time: 2h15m A.1. (DB2_I) Concurrency control (4 points) Page PagA in the database contains tuples t1 and t2. Consider the schedule: r 1( PagA ), w2( t1 ), w1( t2 ) The system relies on hierarchical locking. Show the sequence of lock, unlock, lock escalation and lock downgrade requests for transactions T1 and T2, taking into account that the schedule has to be 2PL. A.2. (DB2_I) Transaction Recovery (3 points) Explain why the warm restart recovery processes redos and undos in two separate phases. B.1. (DB2_II) Distributed Databases (5 points) A distributed DB describes museums in Italy and sup ports the booking of museum tickets for users regis tered on the system. Each museum has a variety of ticket prices and each reservation shows, together with th e museum identifier and the visitor id, the date of t he booked visit and the ticket price. Suppose that the administrative capital of each region contains the description of all the museums in the region, and t hat the overall data about museums (but not the ticket rese rvations) are completely replicated over the DBMSs in Milan, Rome, and Naples. Define the relational sche ma and its distribution. Write at the fragmentation and language transparency levels the SQL statement that moves the “International Museum of Modern Art” fro m Verona to Bergamo (from Veneto to Lombardy). B.2. (DB2_II) Distributed Databases (2 points) Illustrate the behavior of the heuristic recovery in the X-Open DTP protocol. C. (DB2_III) XML (6 points) <!ELEMENT Cocktails ( Cocktail+, Ingredient+ )> <!ELEMENT Cocktail ( Name, Component+, GlassType, Garnish, Procedure )> <!ELEMENT Component ( IngredientName, Quantity )> <!ELEMENT Ingredient ( Name, CaloriesPerGram )> The above DTD describes a…