Redis is categorized as a high-performance key-value database that has become an essential part of big data applications. It supports enterprises, making sense of data by making database-scaling more convenient and cost-effective. 

Having that said about Redis, we know Hadoop is a distributed computing platform. It has high availability, expandability, fault-tolerance, and low operating costs. Hadoop's HDFS storage system makes it difficult to face end user applications. So, computing results are sent offline to user-facing storage such as Redis. 

Horizontal scalability refers to the ability of a system to achieve scalability when the number of machines it operates on is increased. Redis permits data partitioning through range partitioning and hash partitioning. Suppose we have 10 machines, and we pick the key and use the hash function to get back a number. We represent the number, module 10, and the result, in this case 2, is the machine to which the record will be allocated. Replication is accomplished in Redis through master-slave mode. Clients read from the slaves to scale up the read performance. The replication process is synchronous.