In this paper, we proposed a novel scheme, named as TCP-ABC, which replicates the server side TCP connections among multiple server nodes of a cluster. By guaranteeing atomic request delivery, and consensus on responses, this scheme provides the legacy server applications running on the server nodes with multiple active backups in a transparent fashion. And by failing the connections over healthy units, the scheme enhances the service and data availability of the cluster. By conducting experiments on the prototype system on a cluster up to four nodes, we found TCP-ABC results in small performance lost while greatly enhance the service and data availability.