I would like to write an article for this. I recently had to address a real world issue regarding latching and blocking that was caused by parallelism resulting from poor indexing on tables. Things that would be covered in this article:
-Index scans vs seeking
-MaxDOP/Threads/Parallelism
-Wait Statistics
-Graphical execution plans in SSMS
-DMV's to find cached sprocs using specific indexes, amount of times a sproc executed
-DMV's to find indexes on tables, how many times indexes are utilized in execution plans, and seeks/updates/scans per index
The items listed above would be covered in respect to ID'ing parallelism, diagnosing it as the bottleneck, and remedying it with indexes.
Thanks
David Fundakowski