question

Without the ability to see the actual queries, suggesting how to solve the deadlock itself will be very difficult. Remember that deadlocks at their base are performance issues. Despite the fact that evidently two different queries are written in such a way that they're accessing data in different orders, the most common cause of deadlocks, the main error occurs because the queries are too slow. The single best approach to fixing deadlocks is fixing the code. I know you say you can't change it, but that means you're going to have a harder time addressing the core issue. You can look at the queries and see if it's possible to put an index in place to speed things up. You can also look at instituting snapshot isolation as a means to reduce locking on the system. But you probably need to get the code adjusted.