(a) Design a query plan for the following distributed query: An application at site B wants to compute a join of STUDENT ?Id=StudId TRANSCRIPT where STUDENT (Id,Major) is at site B and TRANSCRIPT (StudId, CrsCode) is at site C. The result should be returned to B. Assume that semijoin is not used. Also assume that
- Id and StudId are 8 bytes long;
- Major is 3 bytes long;
- CrsCode is 6 bytes long.
- STUDENT has 15,000 tuples.
- 6,000 students are registered for at least one course. On the average, each student is registered for 5 courses.
(b) Estimate the cost of computing σMajor='CS' (STUDENT) ?Id = StudId TRANSCRIPT using the semijoin strategy. Use the sizes of the STUDENT and TRANSCRIPT relations and of their attributes from the part (a) above. In addition, assume that 10% of the students major in CS. Compare this with the solutions in part (a) above that do not use the semijoin.