1. We've seen how to estimate the size of a join. Now we look at estimating the cost of processing a join.

2. Several factors in fluence the selection of an optimal strategy:

• Physical order of tuples in a relation.
• Presence of indices and type of index (clustering or not).
• Cost of computing a temporary index for the sole purpose of processing one query.

3. We'll look at computing the expression deposit ∝ customer assuming no indices exist. We also let

ndeposit = 10,000 (number of deposit tuples)
ncustomer = 200 (number of customer tuples)