The Falcon engine supplies full ACID (atomic, consistent, isolated, durable) transaction control coupled with crash recovery so no critical data is ever lost.
Falcon manages all transactional activity in memory (with optional paging to disk if necessary) so all transactions are serviced in the quickest time possible. In addition, rollbacks of even the largest transactions are accomplished in a near instantaneous manner.
Falcon's MVCC and row-level concurrency design mean readers don't block writers and vice-versa, with the end result being that lock contention issues are extremely rare.
Falcon utilizes tablespaces for user data storage with there being no practical limit to how many tablespaces can be created and used to manage tables, indexes, and BLOB data. All tablespaces feature auto-extending datafiles, automatic space reclamation, and compaction of data pages.
Falcon sports a distinct memory cache called the Record Cache that differs from traditional database caches in that it only holds the necessary rows being requested by end user queries instead of full database pages that may or may not have data being asked for via SQL queries. This equates to more efficient memory utilization plus it ensures needed data stays hot/resident in memory all the time.
Falcon utilizes special B-tree indexes that always read in physical disk order, which minimizes I/O response time. Multiple indexes can be created on a Falcon table with all indexes servicing requests in this fashion. Indexes also have prefix and suffix compression to reduce I/O load further.
Troubleshooting and tuning Falcon performance is easy via a number of diagnostic tables that provide a window into memory usage, I/O performance, transactional activity, lock activity, and much more.
There is no complexity whatsoever in terms of configuration as only a handful of variables exist to control the behavior of the Falcon engine.
Extreme degrees of high availability are easily accomplished for a Falcon-driven system by using either MySQL replication or supported third-party high availability solutions such as DRBD.
Falcon's design takes advantage of multi-core systems to provide parallel execution of user and service threads. Falcon uses fine-grained multi-threading to increase parallelism with locking on internal structures being done at a low level. In some cases, two threads can change different attributes of the structure at once, because the attributes are separately lockable.

