Exercise 7: Parallel Query Processing.

Consider the following schema

Product(maker, model, type) 
Laptop(model, speed, ram, hd, screen, price)

and the following query

SELECT model
FROM Product P, Laptop L
WHERE P.model=L.model AND P.maker='A'
      AND L.price<1000 AND L.ram=8

Suppose we run the query on a 2-node parallel DBMS where

• Product is partitioned into P1 U P2 on model using hashing
• Laptop is partitioned into L1 U L2 on price using ranges

a) Write the RA for the query

b) Re-write the RA as a union of queries on each of the 2 nodes

c) Under what conditions might some of the queries on the cluster nodes be eliminated? How could those conditions be determined?