The ALICE Detector Control System (DCS) is in charge of control and operation of one of the large high energy physics experiments at CERN in Geneva. The DCS design which started in 2000 was partly inspired by the control systems of the previous generation of HEP experiments at the LEP accelerator at CERN. However, the scale of the LHC experiments and the use of modern, "intelligent" hardware and the harsh operational environment led to an innovative system design. The overall architecture has been largely based on commercial products like PVSS SCADA system and OPC servers extended by frameworks. Windows has been chosen as the operating system platform for the core systems and Linux for the front-end devices. The concept of finite state machines has been deeply integrated into the system design and the design principles have been optimized and adapted to the expected operational needs. The ALICE DCS was designed, prototyped, and developed at a time when no experience with systems of similar scale and complexity existed. At the time of its implementation the detector hardware was not yet available and tests were performed only with partial detector installations. In this paper we analyse how well the original requirements and expectations set ten years ago comply with the real experiment needs after two years of operation. We provide an overview of system performance, reliability and scalability. Based on this experience we assess the need for future system enhancements to take place during the LHC technical stop in 2013.