Datacenter applications frequently cause incast congestion, which degrades both flow completion times of short flows and throughput of long flows. Many congestion control schemes have been proposed to meet the datacenters' performance requirement. Data Center TCP (DCTCP), using Explicit Congestion Notification (ECN) to infer congestion, has been proposed as a TCP variant for current state-of-the-art datacenters and recently has been a standard of IETF. In datacenter networks, where round trip times are in hundreds of microseconds, accurate feedback and quick rate control are crucial to achieve both high utilization and low queuing delay. However, DCTCP loses performance due to its slow and cautious reaction to incast. In this paper, we improve DCTCP and propose a fast congestion control mechanism, called FDCTCP, which can quickly detect incast using ECN, back off according to the congestion gradient and restore previous sending rate once incast ends. Our simulations show that FDCTCP greatly improves performance under different buffer size and significantly reduces the flow completion times, especially, for small and medium size flows compared to DCTCP.