The Compact Muon Solenoid (CMS) experiment is a general purpose detector at the Large Hadron Collider. It has been designed and optimised to discover the Higgs boson and physics beyond the Standard Model. An early discovery of the Higgs boson is the collaboration's top priority and will require a good understanding of both the detector and the physics of the background processes, with a small integrated luminosity. This principle has been the driving force behind the work presented in this thesis. The Silicon Strip Tracker (SST) sits at the heart of the CMS detector. The development of core algorithms to commission the SST are reviewed and the process of live commissioning at the Tracker Integration Facility is described. A crowning success of this study is the calibration of 1.6M channels and their synchronisation to a cosmic muon trigger to within 1 ns. The SST is expected to produce five times more zero-suppressed data than any other CMS subdetector. As such its efficient handling within with High Level Trigger algorithms is paramount. The performance of the online hit reconstruction software is profiled, the inefficiencies are characterised and a new schema to focus on physics regions-of-interest only is proposed. As an example of its success, when running the single 't trigger path over n' -+ r VT events, hit reconstruction times were reduced from 838 ± 5 ms to only 5.13 ± 0.05 ms without any loss in tracking efficiency. The new software is now the tracker community's permanent online solution and is expected to become the offline solution in the near future. bbZO production at the LHC is of great interest, primarily due to its status as a background to a supersymmetric Higgs boson production process. The preparation for a cross section measurement with 100 pb-I of data (expected by the end of 2009) is made. The prominent backgrounds are identified and a signal selection strategy is developed and optimised using Monte Carlo. This study demonstrates that a cross section measurement with this amount of data is feasible. Finally, a method to estimate background from data is tested.