Raid SuperCharger Prevents Spaghetti
Even though SuperCharger writes in linear clusters, it always writes FIFO -in the order data is received - and does NOT reorder data. As a result, it is impossible to have the spaghetti confusion of databases that can be caused by methods such as ladder sorts.
The only question in a system failure in such an environment is whether you need every piece of data available at the instant of failure, or whether you can afford to lose part of a second of data. In most instances, a linearly consistent database is sufficient. Here, you can set the forced flush delay. The range is 3 to 1,000 milliseconds, with a default of 500 milliseconds.
But if you need full TPC compliance, we have a battery or Flash backed up RAM solution which assures that all acknowledged but uncommitted data can be recovered. Normally, TCP compliant solutions are only available through our licensees due to the complexity of implementation and proprietary hardware, so please contact us for sources if you need this feature.