From Rob's comment about a runaway script impacting all connected users (
http://forums.wolflair.com/showpost....2&postcount=28) and the outage/update posts, it sounds like HLO runs on a single server or a very small number of servers. I've been managing servers for over 30 years and I've seen what he's talking about many times.
There are several ways to address something like this but they're all expensive, probably beyond LW's budget - microservers or containerizing the app (Docker or similar) where each user gets their own server/container instance when they connect. It could also run as a cloud-based serverless app, again each connection spawning a separate instance. These are also more difficult to manage and support than a "traditional" web app so add more cost.
It might be possible to monitor each user's processor and memory consumption and terminate processes or connection that exceed a threshold, it depends on how HLO is implemented and kind of monitoring is available to LW.