The biggest limiting factor is that the sync engine cannot be configured for load balancing or redundancy, followed closely by the through-put limitations that apply to the FIM Service connector (management agent) when dealing with high volumes of objects under synchronization. Nonetheless, FIM has been successfully deployed to manage the entire identity life-cycle for very large sites, such as well over one million staff/student identities under management for the Queensland Education Department in Australia. Also, when the "declarative" sync rules requiring "expected rule entries" (EREs) are employed, the sync overhead is further exacerbated. As such this type of sync approach should be avoided in lieu of a code-based rules extension or the "scoped" sync rule capability released with R2.