receiving a query to be searched in the cache hierarchy, wherein the cache hierarchy comprises a first cache and a second cache,

wherein the first cache comprises a first partition and a second partition,

wherein the first partition is associated with a lock counter,

wherein the lock counter tracks a number of threads attempting to access data in the first partition, and

wherein delay caused by a locking mechanism associated with the first partition is modeled based, in part, on the lock counter;

computing a hit rate value associated with the first cache in the cache hierarchy, wherein the hit rate value is computed based, in part, on a counter group, wherein computing the hit rate value comprises computing a ratio of current contents of the first cache over a size of a file associated with the query, and wherein the file is cached in the first cache;

comparing the hit rate value to a predetermined value to obtain a result, wherein the predetermined value is a number randomly selected from a uniform distribution, wherein a hit is determined when the predetermined value is not greater than the hit rate value;

providing a reply to the query when the result corresponds to the hit;

incrementing a counter in the counter group when the first cache is not full;

proceeding to the second cache in the cache hierarchy when the result is a miss; and

providing a diagnostic output comprising a overall performance for the database management system, wherein the overall performance is determined, in part, on the delay caused by the locking mechanism associated with the first partition.