All vector timestamps have the same number of elements in the vector $($fixed length$).$ The vector timestamp assigned to a transaction is fixed, because it refers to the vector timestamp at the time the when the transaction starts, and no two transactions have the same timestamp. The system currently uses exclusive locking of objects for concurrency control and relies on $2-$phase locking $($as discussed in class$).$ However, the system restricts when a transaction can block $($i$.$e$.,$ wait$)$ when acquiring locks. Specifically, given two transactions Ta and Tb$,$ when Ta is trying to acquire a lock on an object that is currently held by Tb$,$ Ta will abort itself $($and not wait for the lock to be released$)$ if Ta is the younger transaction $($i$.$e$.,$ TS$($Ta$)>$ TS$($Tb$)).$ Otherwise, Ta waits for Tb$.$ They tell you they are facing deadlocks but have no idea why. a$.$ Give a counterexample to show how a deadlock may occur.