This paper brings out the current state of art in multi-hop routing techniques for sensor networks, specifically for many-to-one periodic data collection. We evaluate the routing layers on a large scale Sensornet testbed of 100 nodes where the nodes are densely spaced. The evaluation was focused on trying to short list the routing layer with the best data reliability. Such an evaluation is expected to be useful to researchers building pervasive applications using TinyOS based sensors (micaa2 and mica2dot) and a multi-hop routing stack. We saw that the reliable time sync and the DSDV implementation in HSN (heterogeneous sensor network) routing stack performed the best, with percentage of successful transmissions averaging at 58.3% and 62.9% respectively. The nodes that were closer to the base station had a high success rate of 86.9% for HSN DSDV and 83% for reliable time sync.