Recent research indicates rising patterns in undesirable network traffic, perhaps most markedly in the presence of Peer to Peer file sharing. The resulting burden this traffic places on limited network resources frequently leaves administrators and organizations seeking effective network management tools. The packet shaping tools available to administrators are typically expensive, closed source, black-box technologies that provide customers with minimal insight into the internal techniques used to manage their traffic. The aim of my submission is to provide the foundation for an open source alternative to costly commercial technologies that can rival and, potentially, outperform these shapers on comparatively low end hardware. A central component of my proposed system relies on eliminating the need to closely scrutinize and filter each packet captured by the shaper. The replacement for this mechanism relies on a low overhead system capable of quickly decomposing packet headers and maintaining general data on streams. The system uses these statistics to target the most bandwidth intensive and most likely P2P traffic streams, along with a comparatively small number of randomly selected streams. The targeted streams are then submitted for low level examination via a series of filters, similar to those that might be found in a standard shaper. The filters will provide a final verification of a stream's validity and may result in the stream's elimination. A further layer will provide the ability to bind traffic to a user's identity, allowing misbehaving users to be identified and allowing all streams for that user to be immediately examined and terminated over longer periods of time if necessary.