This paper summarizes discoveries and remedies of simulation protocol data loss across the Battle Lab Collaborative Simulation Environment (BLCSE) Wide Area Network (WAN). The use case for a network supporting large-scale constructive simulation, combined with other traffic, carries special requirements outside the typical boundaries for normal systems administration of a WAN. Understanding the challenges and solutions involved is certainly not mundane. Simulation data losses in excess of 1% can impose compounded causal effects that will easily jeopardize the analytical benefit of an experiment. The possibility of data loss must be vigilantly monitored and vigorously interdicted. One root cause of packet loss is network congestion. Congestion occurs at chokepoints, which exist in nearly all network topologies where a number of hosts on a local network aim to connect to resources at remote destinations via a shared infrastructure. Another, unanticipated cause of loss is technology integration conflict, based on original design assumptions. Finally a more surprising and insidious cause is the burstiness of simulation traffic in which High Level Architecture (HLA) packet loss occurs at utilization levels far below the bandwidth threshold (i. e. without congestion). Over time, the approach of the Cyber Enterprise Support Center (CESC) and Army Capabilities Integration Center (ARCIC)/Joint Modeling and Simulation Divisions' (JAMSD) management of the BLCSE network has evolved from reaction, to monitoring, to deliberate stimulation and most recently to the intentional, governing configuration and application of the Cisco IOS Quality of Service (QoS) technology. The hard-felt experiences of the BLCSE community, as well as the powerful off-the-shelf and custom technologies used will provide a tremendous value to the greater modeling and simulation community, and should be seriously considered for other wide-area simulation environments.