An application proxy firewall is able to scan all incoming application data for viruses. It would be more efficient to have each host scan the application data it receives for viruses, since this would effectively distribute the workload among the hosts. Why might it still be preferable to have the application proxy perform this function?